@science.org warns against AI slop in research: “authentic research and validated findings will slow down to an expensive trickle into established knowledge, and the opportunities for flawed or fabricated information to be seen as reality will grow” www.science.org/doi/10.1126…
Who's Who of AI
Claes de Vreese
Why they matter
Researcher with public evidence across AI research, AI business.
- AI signals
- 7 past 30d
- Sources
- 7 distinct domains
- Discussions
- 0 past 30d
- Latest signal
- 2d ago
Professor of AI, society, media and democracy + Political Communication, U Amsterdam || Director of Digital Democracy Centre, SDU || AlgoSoc || AIMD || Mix of work and private stuff ||
🇩🇰🇳🇱🇪🇺🏃🏼🧪☕️🍿🐶🟥🦋
What they're sharing
science.org
Landmark German ruling declares Google's AI Overviews are Google's own words and makes it liable for false answers the-decoder.com
Language Models Trained on State Media Sources Launder Propaganda | TechPolicy.Press techpolicy.press
reuters.com
Europe’s 27 leaders to take on AI — finally politico.eu
openai.com
Commission publishes Code of Practice on marking and labelling AI-generated content digital-strategy.ec.europa.eu
OpenAI says its models went rogue and hacked startup in ‘unprecedented incident’ theguardian.com
Governments Want to Ease AI Regulation for Innovation, But Do Citizens Agree? techpolicy.press
American Democracy Wasn’t Designed for This theatlantic.com
Waarom AI-bedrijven je bang willen maken: 'We denken al snel aan The Terminator' nos.nl
View on Bluesky →
The Authenticity Challenge | Weizenbaum Journal of the Digital Society ojs.weizenbaum-institut.de
Articles & links
Really interesting. Liable for content! The court treated the AI overviews as Google's own content and rejected Google's argument that users were responsible for fact-checking the results themselves the-decoder.com/landmark-ger...
AI Weekly's analysis
→
- Munich's Regional Court declared Google's AI Overviews its own speech, rejecting traditional search-engine hosting liability protections.
- Google's AI Overviews answered correctly around 91% of queries, but 56% of those correct answers couldn't be verified through the linked sources.
- Google confirmed it will appeal the ruling, which could set precedent for AI answer engines including ChatGPT and Perplexity across Europe.
Read full analysis →
Training data for AI models matter for what you get. A lot. And states can benefit. Good primer by @justinhendrix.bsky.social and @techpolicypress.bsky.social on @hwaight.bsky.social and team’s important piece in @nature.com www.techpolicy.press/language-mod...
Pope Leo warns about AI. A warning against misinformation, conflict, and the risk of “leading the world down a path of unending war” Pope Leo read the room right at least on one thing: citizens support regulation of AI Read: www.techpolicy.press/governments-... www.reuters.com…
This story credits Member of the European Parliament @reiniervanlanschot.volteuropa.org to call on Costa to "convene a special European Council summit dedicated to artificial intelligence as soon as possible." www.politico.eu/article/euro...
Open AI says it will support the EU Code of Practice on Transparency of AI-generated content. A good start. Now let’s move from the announcement to signing the Code. openai.com/index/suppor...
The 🇪🇺 Code of Practice on marking and labelling AI-generated content has dropped💥 The Code sets out practical measures for providers and deployers of Gen AI systems to comply with the AI Act’s transparency obligations. They apply from August 2nd 2026 digital-strategy.ec.europ…
AI Weekly's analysis
→
- The Code of Practice is voluntary, but the AI Act transparency obligations it supports become enforceable on August 2, 2026.
- Three categories require labelling: deepfakes, AI-generated or manipulated text on matters of public interest, and interactive AI systems such as chatbots.
- The European Commission released EU icons for labelling AI-generated content alongside the code to reduce implementation friction for providers.
Read full analysis →
“The company behind ChatGPT said Hugging Face had detected and contained the agent” AI agents should *not* be described as human-like. They are not. www.theguardian.com/technology/2...
Pope Leo warns about AI. A warning against misinformation, conflict, and the risk of “leading the world down a path of unending war” Pope Leo read the room right at least on one thing: citizens support regulation of AI Read: www.techpolicy.press/governments-... www.reuters.com…
The challenges of new (communication) technologies for democratic institutions, discussed with the United States as a backdrop. Of course the challenges are in no way uniquely American. But the technology owners almost all are. www.theatlantic.com/ideas/2026/0...
Claes de Vreese reposted
Sprak met NOS over imaginaries van AI. En de hardnekkigheid ervan. Het feit dat de beeldredactie dan alsnog een foto van terminators kiest, is denk ik het treffende bewijs daarvan 😅🥴 nos.nl/artikel/2624...
🚨 new publication with @nickmattis.bsky.social: “The Authenticity Challenge” On the transformative nature of generative AI for journalism and the need to redefine authenticity as a process to preserve credibility in an AI saturated information ecosystem. ojs.weizenbaum-institu…
Their own posts
Recent commentary
US president using AI to mock and brag about his censorship efforts of media/ satire/ free speech, amplified by official White House communication channels. All normal stuff.
It’s hard: journalists are not that good at detecting AI content. Luckily they are better at detecting real content. But still … @lyngemoller.bsky.social presenting our (with @emilbakke.bsky.social and 🙋) paper #ica26 #awardwinning 😀 @ddc-sdu.bsky.social
It is recommendable that the new Danish government program has a full chapter on AI and an ambition to offer a balanced perspective. But there are a LOT of things to work out, many open ends. The devil is in the details. Hopefully they will leverage expertise and research in the next steps #dkpol
Their network
In Claes de Vreese's orbit
Center = Claes de Vreese. Left = members they follow (green edges). Right = members who follow them (blue edges). Top = mutual follows (orange edges, slightly larger). Drag any node to reposition; click to open that profile.