AI models are using material from retracted scientific papers www.technologyreview.com/2025/09/23/1...

"In a landmark study, OpenAI researchers reveal that large language models will always produce plausible but false outputs, even with perfect data, due to fundamental statistical and computational limits." www.computerworld.com/article/4059...

Aaron Parnas debuted on TikTok posting about pizza toppings and hair loss. Now, at 26, he's pivoted into journalism, landing interviews with leading Dems and becoming one of Substack's most popular creators—while accepting dark money. New @columjournreview.bsky.social www.cjr.org/feature/aaro...

Generative AI and the German Far Right: Narratives, Tactics and Digital Strategies www.isdglobal.org/isd-publicat...

Political Advertising on Social Media EDMO Belux's recent study explores these big questions: How will these differences impact the future of political advertising online? And what does it mean for the transparency of these platforms? Read more: buff.ly/1hkpwRZ #SocialMedia #PoliticalAdvertising

Universities and their communities play an essential role in protecting and promoting academic freedom. 🛡️ 🤲 They must commit to making academic freedom a lived and daily reality. Read EUA's position paper buff.ly/35qwaif #AcademicFreedom

Toronto indie mag pushes back against ‘AI-generated slop’. Editors at The Grind caught seven articles for one special edition they suspected were AI fakes. j-source.ca/toronto-indi...

Skeptik: A hybrid framework for combating potential misinformation in journalism, "often stemming from flawed reasoning and logical fallacies" and "poses significant challenges to public understanding and trust in news media." arxiv.org/abs/2508.18499

"Universities must take their role seriously to a) counter the technology industry's marketing, hype, and harm; and to b) safeguard (...) critical thinking, expertise (...) and scientific integrity. " in Against the Uncritical Adoption of 'AI' Technologies in Academia zenodo.org/records/1706...

Le fait émergent, ou (re)penser ce qui fait le fait - réflexion autour de mes travaux sur la factualité à l'heure des IA génératives à l'Université de Bergen. www.meta-media.fr/2025/08/01/l...

Vi sees neste år!

Ravie d'avoir participé au dernier n° de la revue I2D @adbs-infodoc.bsky.social "Désinformation, véracité et traçabilité de l'information" avec une analyse sur le journalisme et le devoir de vérité à l’épreuve de la désinformation et des contenus générés. shs.cairn.info/revue-i2d-in...

📣AI propaganda factories🏭 are now operational. My study shows how small, open-weight models can run as fully automatic generators in influence campaigns. mechanising personas, engagement, cadence. Possible for State, non-state, and micro-actors, including and bedroom ones.

History is always so repeating! Here is a paper by Woolgar, published in 1985: Why not a Sociology of Machines? The Case of Sociology and Artificial Intelligence journals.sagepub.com/doi/abs/10.1...

Don't use Copilot für anything with numbers in Excel ,)

TOP DOWNLOADED ARTICLE! We’re happy to see so much traction for our commentary on defending #academic #freedom amid increasing authoritarianism in democratic societies. This article, penned by 13 senior scholars led by @oscarwestlund.bsky.social, is available for FREE ACCESS here: bit.ly/47FH7M7.

The commentary is also seeking signatures from researchers from all countries to support this important cause of defending #academic #freedom! To be a signatory, please click on this form: bit.ly/474kO2D. We thank you for your support!

The only thing we learn from the history of tech bubbles is that we learn nothing from it - Why the AI Hype is Another Tech Bubble link.springer.com/article/10.1...

the AI hype bubble is coming. And the news organizations and pundits that helped inflate it are going to immediately pivot to pretending they saw it coming the whole time.

Marre des sites d'information générés par IA (GenAI) ? Notre extension (gratuite, pour Chrome & Firefox) affichant un message d'alerte sur les sites GenAI (en français) en répertorie désormais plus de... 6 000 (contre 1 000 en février). Faites tourner ! next.ink/195816/alert...

📍[Publication] Comment distinguer le vrai du faux, comment développer des méthodes pour détecter la désinformation dans nos structures professionnelles ? Le numéro de rentrée de la revue de l’ADBS 👫👭 arrive ! 📆 Parution le 19 août, avec à la une notre dossier spécial #désinformation.

the ai-powered pdf is a harbinger of doom www.wired.com/story/adobe-...

🇸🇪 Chez Aftonbladet, le grand tabloïd suédois, le ChatBot maison « Hej Aftonbladet » rencontre visiblement son public - 50'000 questions posées par jour. Il faut dire que l'expérience utilisateur est plutôt assez propre. ⏩ Lire : www.linkedin.com/posts/martin... ⏩ Le tester www.aftonbladet.se/hej

‘We didn’t vote for ChatGPT’: Swedish PM under fire for using AI in role www.theguardian.com/technology/2...

A linguistic engineering approach to prompting #NordMedia #NordMedia25 @nordishub.bsky.social ohmybox.info/a-linguistic...

An important piece on #AI generated #preprints and why it's important to consider their impact on scientific ecosystem. Some other important considerations related to this that are not covered in the article 👇 🧵1/14 #AI4Science #AI4Research #ScAISci #SciSci

In which a cognitive scientist discovers that the definition for ‘domain-general cognitive ability’ on *ScienceDirect* is AI-generated slop, and completely wrong

🟢 We are looking for journalists reporting on #climate #disinformation 📝 Apply if you have research ideas on the threat of climate denialism, the relation between big tech & climate obstruction or similar 👇 eu.boell.org/en/climate-d... ⏳ Deadline: 9 Sept 2025, 23:59 CEST @zorasiebert.bsky.social

Publié dans la dernière livraison des Cahiers de Tendances de France TV, cet article s’appuie sur une recherche menée sur le concept de factualité à l’Université de Bergen, dans le cadre de @nordishub.bsky.social "Le fait émergent, ou (re)penser ce qui fait le fait" www.meta-media.fr/2025/08/01/l...

New eye-tracking study showed that Haredi participants paid less attention to key credibility cues like metadata (e.g., author bio, source credibility cues), which correlated with lower accuracy in spotting misinformation. www.nature.com/articles/s41...

I'm sure this is a serious social issue (on which I'm not qualified to comment) but I did find this genuinely hilarious: 'The company is yet to experiment with creating plus-size women, claiming "the technology is not advanced enough for that".'

Are you exploring #AI issues and ethics in journalism? We have two pages full of resources for you on The Journalist's Toolbox. Guides and article: journaliststoolbox.ai/ai-tools-and... Ethics/law: journaliststoolbox.ai/ai-and-ethics/

Utilizing AI Text Checkers: Case Studies and Implications www.scirp.org/journal/pape...

Between fact and fairy: tracing the hallucination metaphor in AI discours link.springer.com/article/10.1...

Is RRF the Secret to Dominating AI Citations? I Decoded ChatGPT’s Ranking Formula metehan.ai/blog/chatgpt-is-using-reciprocal-rank-fusion-rrf/

🗣️ Excited to share our new #ACL2025 Findings paper: “Just Put a Human in the Loop? Investigating LLM-Assisted Annotation for Subjective Tasks” with Jad Kabbara and Deb Roy. Arxiv: arxiv.org/abs/2507.15821 Read about our findings ⤵️

Um… any privacy issues here? “It may prove hard to resist an artificial companion that knows everything about you, never forgets, and anticipates your needs better than any human could.” www.newyorker.com/magazine/202... A.I. Is About to Solve Loneliness. That’s a Problem | The New Yorker #eduskyAI

Just from a quick look, this new study from Kobi & colleagues is mindblowing. - Conversational LLMs are persuasive & they are so because they are good at using facts and evidence that support their arguments - the information used to persuade was mostly accurate

On platform dependencies: Google kills the fact-checking snippet www.niemanlab.org/2025/06/goog... The last IFCN newsletter highlights that "More than 250,000 fact checks have been tagged with ClaimReview, and Google itself reported in 2019 that annual impressions likely reached 4 billion."

A major AI training data set contains millions of examples of personal data www.technologyreview.com/2025/07/18/1...

It sometimes takes a while for certain news to reach us. I just learned about Bartosz Wilczek. He was a great researcher and a truly kind person. So sad.

AI in journalism: Some perspectives from the Global South #academicreadings ohmybox.info/ai-in-journa...

Falling for Russian Propaganda: Understanding the Factors that Contribute to Belief in Pro-Kremlin Disinformation on Social Media journals.sagepub.com/doi/10.1177/...

Is Objectivity in Journalism Still Worth Pursuing? By Julie Gerstein and @sulliview.bsky.social‬ of @newmarkjresearch.bsky.social‬ at @columbiajournalism.bsky.social. www.cjr.org/analysis/is-...

🚨 New publication alert 🚨 Aqsa Farooq and I explore how aesthetic realism and emotional salience influence authenticity judgements of AI-generated disinformation images. #GenAI #disinformation #authenticity #commsky link.springer.com/article/10.1...

Wow — banning AI tools in coding actually increased performance among experienced programmers, according to their own perception as reported in the survey. They believed they were 20% more productive, while in reality, their performance dropped by 19%? metr.org/Early_2025_A...

Mapping Global Emerging Scholarly Research and Practices of AI-supported Fact-Checking Tools in Journalism www.tandfonline.com/doi/full/10....

Caught in a bind? Some thoughts about academic writing in English as a non-native language: 1) We tend to reproduce the patterns of sentence structures we read, so an AI-text checker might flag a text as AI-generated when it is not. 1/3

Why does a social media company want an AI chatbot? If that AI is then positioned as an arbiter of truth, as in the trend of asking Grok to weigh in on matters, then one now has enormous control over what is “true”. This is somewhere between Bond villain shit and Mission: Impossible villain shit

Public Knowledge and Expertise Under Authoritarian Siege: A Defense of Academic Freedom from Digital Journalism Studies - "This commentary examines the impact of political interference, censorship, and self-censorship in academia" www.tandfonline.com/doi/full/10....