"ChatGPT, the most heavily used AI service, gave wrong information in 46% of its answers, including making up an expenses scandal, giving inaccurate replies on voter eligibility rules and getting the date of the election wrong by two months." www.theguardian.com/technology/2...

Particularly in the context of (appropriate) crackdowns on LLM-generated papers (e.g., on arXiv) and more and more tales of fabricated citations, two reminders: (1) LLMs do not "cite their sources" (2) Using an LLM to generate, complete, or fix LaTeX code can result in fabricated citations

Weird … the Xitter account that I haven’t touched since late 2022 was just suspended for “inauthentic behaviour”. Anyone else experiencing similar? I suspect someone reported it for something, but of course I’ve got no idea what.

The full AAMAS 2026 program has just been released—check out what’s in store for this exciting event. 🗓️ 📍 cyprusconferences.org/aamas2026/de... #AAMAS2026 #Cyprus #AutonomousAgents #MultiAgent

Wise comments from my colleagues — a wafer thin purported justification makes the slated bar on tort claims a worrying breach of the rule of law and democratic traditions etc. theconversation.com/changing-cli...

Mighty congratulations Charlotte + @bwbnz.bsky.social 👏 arts.unimelb.edu.au/news/charlot... and a reminder that this was a @royalsocietynz.bsky.social Marsden funded project in Humanities! #HumanitiesMatter #history @vicuniwgtn.bsky.social

a requirement that subsequent arXiv submissions must first be accepted at a reputable peer-reviewed venue.

We are implementing a similar policy at @arxiv.bsky.social. If there is incontrovertible evidence of LLM slop in a paper, this means the authors did not take the time to read the LLM output and we can't trust anything else in the paper. Penalty is 1 year ban from arXiv followed by...

Really important tidbit here as well: even asking AI to *format* citations can result in hallucination. (I know this but my students absolutely do not believe it)

a small twist on a classic puzzle aifails.substack.com/p/backpack-w...

Not only does telling a chatbot ‘do not hallucinate’ misunderstand how LLMs works, the idea that they are capable of following instructions is an illusion. At best, this just adds text to the context that may cause them to go down one path rather than another. But they don’t literally obey commands.

🤦‍♂️

This is huge. I can't believe it took this long before accountability was established in a domain so evident and hazardous as traffic. But proves its doable. Now let's see it enforced, rapidly and at appropriate scale to effect change. #AIEthics #driverless

The countdown has begun! In just 20 days, we’ll gather for the 25th International Conference on Autonomous Agents and Multiagent Systems. 🔗 Registration System: easyconferences.org 🔗 Registration Guidelines: cyprusconferences.org/online-regis...

Under California's new rules, police can cite AV companies when their vehicles commit moving violations. The rules will also require the companies to respond to calls from police and other emergency officials within 30 seconds. www.bbc.com/news/article...

Babe, wake up, new anti surveillance tech just dropped…

I’m attending, but online, not in person.

Thank you to all of you who attended ACM's Technology Policy Council's special session, "From Technical Safety to Societal Impact: Rethinking AI Governance," at the recent India AI Impact Summit in New Delhi! If you missed it, here's a recap: buff.ly/MDl8s6T

What is the one thing that most worries you about AI? In a survey of 3,700 AI researchers, only 3% replied existential risk—despite the prominence given to these risks in media. Far more experts are worried aout malicious use, misinformation, job losses, bias, etc. www.nature.com/articles/d41...

Little known fact: it's only our data that keeps data centers from flying into space. aifails.substack.com/p/data-cente...

We would like to welcome Professor Sabine Hauert from the University of Bristol to the AAMAS 2026 keynote stage! Discover more about her speech on "Designing Trustworthy Robot Swarms for Real-World Applications". ➡️ cyprusconferences.org/aamas2026/ke... #AAMAS2026 #Cyprus #AutonomousAgents

Prepare for an inspiring keynote speech in Autonomous Agents and Multiagent Systems! Joanna Bryson from the Hertie School will deliver a keynote address on "the Limits of Artificial Agency" at AAMAS 2026🎤 More information ➡️ cyprusconferences.org/aamas2026/ke... #AAMAS2026 #Cyprus

Michael Wooldridge from the University of Oxford will share expert insights at the 25th International Conference on Autonomous Agents and Multiagent Systems!🎤 More info ➡️ cyprusconferences.org/aamas2026/ke...

New: Wikipedia has officially banned AI-generated content. After months of heated debate and previous attempts to restrict the use of LLMs, volunteer editors accepted a new policy that prohibits using them to create articles for the online encyclopedia www.404media.co/wikipedia-ba...

TBH I got the answer I deserved

Tell the AI giants to stand together against killer robots actions.eko.org/a/tell-the-a...

I'm commenting here as an expert: "JESUS LOL FUCK NO".

Embedding unreliable AI deep into the source code of the military without human oversight is the single stupidest thing we could do as a species. And that is *exactly* what the United States is about to do.

(1/9) I love Mike Wooldridge's delivery of some of the examples I post here :-) www.youtube.com/watch?v=CyyL... Maybe also a good time to clarify how I myself interpret these examples.

5/5 What I do see as the big challenges for the rest of the 21st century (all my grandchildren can be expected to live beyond the year 2100), are global warming, and rise of diseases due to vaccine deniers. Stay calm. Love those around you. Be kind to everyone.

1/5 I am getting more and more emails from individuals directly to me expressing fear for their children from possibly physical actions from AI. They mention that my views seem balanced and ask me to respond to their words. The attached four skeets are a response to one such that I wrote today.

Ready to register? Visit www.easyconferences.org 🔹 Reserve accommodation 🔹 Purchase extra tickets 🔹 Sign up for workshops 🔹 Print your poster locally 🔹 Arrange airport transfers 🔹 Book optional excursions ✅ For more information contact @easyconferences.bsky.social #AAMAS2026 #Cyprus

The wait is over! 🎊 The AAMAS 2026 Accepted Research Track papers are now live! 🔗 View the full list of accepted research papers here: cyprusconferences.org/aamas2026/ac... Join us in celebrating the hard work of our global research community! #AAMAS2026 #MultiAgentSystems #AIResearch

Join our hands-on workshops at AAMAS 2026 and gain practical skills from leading experts in Autonomous Agents and Multiagent Systems workshops. 🎟️ Learn more: cyprusconferences.org/aamas2026/wo... #AAMAS2026 #Cyprus #AutonomousAgents #MultiAgent

New AIWeirdness post: ChatGPT will apologize for anything, including stuff that didn't happen. It's not reflecting on stuff it did wrong, it's improv. www.aiweirdness.com/chatgpt-will...

We wish you happy new year, and we can't wait to meet you in Cyprus this year! #AAMAS2026

Happy New Year! See you in Cyprus soon! #AAMAS2026 #AAMAS #2026

Happy new year! (My brain hurts from reading this 😵‍💫)

Ban the Business Model, Not the Child. Read these wise (research based) words from colleague @jesstpiotrowski.bsky.social on the social media ban in Australia @ascor.bsky.social @antmandan.bsky.social www.jessicataylorpiotrowski.com/2025/12/11/b...

Researcher produces 100+ papers on "AI" in a year and what's worse than the shonky vibe scholarship and "slop flood" this represents is that he's doing it all as boss of an edtech company called Algoverse charging students $$thousands to do the actual work www.theguardian.com/technology/2...

From a small, one-day event in 2023, to this top-speaker stacked two-day conference we got now. Really excited to see the programme of the Capital City Complex Systems Symposium to be held end of February 2026 published. Register now and see you in Wellington.

Happy Thanksgiving everyone!

In a sort-of social experiment, instead of assigning "reviewers" to the submitted papers, we assigned "mentors". Although the job content was the same, the quality of the reviews and the satisfaction of the authors improved quite significantly (even those whose paper was rejected). Think about it.

It’s widely known (and, I think, pretty uncontroversial) that learning requires effort — specifically, if you don’t have to work at getting the knowledge, it won’t stick. Even if an LLM could be trusted to give you correct information 100% of the time, it would be an inferior method of learning it.

instructions for air travel

My views on "agentic" here in this AP article on "agentic AI" apnews.com/article/agen...