I just added to my personal website that "I never have and never will work for big tech, and I have never signed a non-disparagement agreement." stephencasper.com/

I really liked this video: youtu.be/UgDEyQ1h-EA...

In practice, almost all AI porn is, in a certain morally relevant sense, not completely consensual. I was glad to talk briefly about this with Jessica Klein while she was writing this new MIT Technology Review piece!

As an AC for the ICML Technical AI Governance Research workshop, I am noticing a trend. A LOT of papers have the word "increasingly" in the first sentence of the abstract. 🙃

🧵It is hard to overstate how disappointing I think this new paper from Oxford, OpenAI, Anthropic, and Google (et al) is. I can't take it seriously as academic work, just as propaganda. It also has some very bad scholarship and questionable adherence to research ethics.

Governing AI is kind of like grief: 1. Denial -- "...next token prediction...stochastic parrots..." 2. Anger -- "Elon did WHAT?!" 3. Bargaining -- "Maybe Newsom will sign a watered-down bill." 4. Depression -- "Oh, this is what Collingridge was talking about." 5. Acceptance -- :/

When talking about AI and its future, I wish it were more of a faux pas to make vague appeals to the benefits of technological progress -- and I don't just mean "AI curing cancer". Technological progress in general probably isn't as good as we normally think it is.

A recent (short) talk of mine is now up on YouTube. It is about why, if you are working for an AI company on making its systems safer, I think you should consider quitting. www.youtube.com/watch?v=Ugu7...

🚨 One week left to submit your AI-gov-related research to the TAIGR workshop.

Now that Mythos is released, we can start the clock. I'd bet that within 9 months, a system with comparable cyber capabilities will be widely available (either open-weight or openly served). Hopefully, we just have enough time to improve cyberdefense enough to be ready.

Now that Mythos is released, we can start the clock. I'd bet that within 9 months, a system with comparable cyber capabilities will be widely available (either open-weight or openly served). Hopefully, we just have enough time to improve cyberdefense enough to be ready.

🧵🧵🧵 A provocation to the mechanistic interpretability researchers of the world...

Me in December: "Wow, freedom of information laws are awesome. I can't wait to get info." Me now: Enters my 4th month of getting bullied and gaslit by 3 governments at once. Anyway, if you need to do US FOIA, UK FOI, or EU FOIA requests, let me know -- I have advice.

Reasons to submit to the ICML Technical AI Gov. Research (TAIGR) workshop: - 8-page limit - Broad scope, AI gov-related - Workshops don't trigger dual submission policies - Best paper awards both overall and by category - Great community - Cool stickers Deadline April 24!

OpenReview for the #TAIGR workshop for #ICML on technical AI governance research is live as of today. (Not an April Fools joke). See the call and link to OpenReview here: taigr-workshop.com

I wish more CS papers had tables of contents. Makes them much more navigable. I think one reason it's rare is that submission venues have page limits. So there's often just not room. I wish venues would conditionally relax length requirements by the length of an optional ToC.

Today, I realized that, taken together, Appendices 3.2, 3.3, and 3.5 of the EU AI Act Codes of Practice *unambiguously* require open-weight model developers to let external evaluators conduct adversarial fine-tuning evals on their frontier models. Good.

Could social media make us less polarized instead of more? We tested 5 algorithms on 3 platforms with 10,000 people for 6 months during the 2024 election, and found that the answer is yes. 🧵

Just like we have mass school shootings, we now also have mass school AI non-consensual nudifications.

Announcing the technical AI Governance Research (TAIGR) ICML workshop in July! Submissions (up to 8 pages) are due April 24. Co-submission with ICML and NeurIPS is encouraged. taigr-workshop.com

Do you do technical AI research? In this talk, I argue that you 🫵 should see yourself quite literally as a type of policymaker. Thanks @far.ai. www.youtube.com/watch?v=Ekp...

The official text isn't out yet & will matter a lot. But Europe may be moving to treat AI nudification similarly to how pirated media or CSAM is treated: as something that, even though it will always be available, can possibly be made much less accessible. www.politico.eu/article/eu-g...

Using a well-timed screenshot and my phone's cache, I was able to recover some of the since-deleted tweet from Jeremy Lewin, where he admitted that the government sees the new OpenAI contract language as just "memorializing" a vague "commitment" rather than drawing any real new lines.

As someone who is not a fan of @anthropic.com...I think you should use Claude.

We almost certainly won't make AI safe by making safe AI. Others are still going to create unsafe AI. – @scasper.bsky.social at #IASEAI2026 Open-Weight AI Risk Management Workshop I led one of the discussion groups and we had some nice new ideas of how to make open weight models safe 😊

Given: 1. Last summer, frontier closed-weight model devs started to share warnings about nasty model capabilities. & 2. Open-weight models are a few months behind closed ones. We should not be surprised if there is a big cyber/terror incident enabled by a powerful open-weight AI model in 2026.

The 2025 AI Agent Index is now available on arXiv. arxiv.org/abs/2602.17753 x.com/StephenLCas...

If you're responding to @facct.bsky.social rebuttals this week, feel free to use my rebuttal template. docs.google.com/document/d/...

🚨The 2025 AI Agent Index is out! 🚨 Amidst recent buzz over 🦀 and NIST's new agent initiative, we find: - Selective reporting – esp. on safety - Almost all agents backend just 3 model families - Many agents don’t ID themselves as bots online - Big US/China gaps - And more…

The International AI Safety Report is now available in all 6 official UN languages. 🎉 internationalaisafetyreport.org/publication...

A project idea that you can feel free to scoop me on -- I'm not working on this (yet?). Lately, when I have talked with AI governance people about non-consensual AI deepfakes and CSAM, the conversation almost always touches on a particular open problem...

🚨 New paper led by Joe Kwon with GovAI. Are you worried about OpenAI automating dev & evals with AI agents? What about Grok reading all of your tweets & info to profile you? Some of the most consequential *internal* deployments of AI systems are in regulatory grey areas.

Did you ever notice that the image at the top of OpenAI's "Our approach to AI Safety" article is a giant red flag???

On the @csis.org podcast with Greg Allen and Stephen Clare, we talk about the International AI Safety Report, technical safeguards, and what engineers can (and can't!) do to save us from AI risks. www.youtube.com/watch?v=2VlX...

Personally, I wish Anthropic would go a step further and also mention at the end of their ads that they aren't currently embroiled in multiple lawsuits over the deaths of children. 🧵🧵🧵 What I like about the new Claude Ads www.theverge.com/ai-artificia...

The 2026 International AI Safety Report has 221 pages of cutting-edge research from a team of 36 writers and over 100 contributors referencing 1451 citations... ...And also this cheeky riddle that I wrote on page 23 (original content). internationalaisafetyreport.org/publication/...

The IAISR is one of a kind. Every paragraph has undergone many rounds of scrutiny from dozens of experts and stakeholders over the course of months. I'm thankful for the rest of the writing team. If you're interested, my work this year was mostly in sections 1.1 and 3.3.

In a few years, I think it would be cool to do a project and write a paper on how prominent AI systems from US companies respond to political questions before vs. after the next presidential transition.

Turns out, there are a TON of image/video AI models hosted on CivitAI with dogwhistles for NCII and/or CSAM in their names. 👀 Max Kamachee and I just updated our "Video Deepfake Abuse" paper with this new fig: 🔗 papers.ssrn.com/sol3/papers....

Open-weight model safety is AI safety in hard mode. Anyone can modify every parameter. @scasper.bsky.social: Open-weight models are only months behind closed models, which are reaching dangerous capability thresholds. 2026 will be critical.👇

This is not a new report (it's from last summer). But it's now finally available on SSRN, more accessibly than before. Great working with Claire Short on this. papers.ssrn.com/sol3/papers....

I made a fully-open, living document with notes and concrete project ideas about tamper-resistance and open-weight model safety research. You, yes you 🫵, should feel free to look, comment, or message me about it. docs.google.com/document/d/1...

Here are some miscellaneous title ideas for papers that I'm not currently working on, but sometimes daydream about. Let me know if you are thinking about anything related.

Research on tamper-resistant machine unlearning is funny. The SOTA, according to papers proposing techniques, is resistance to tens of thousands of adversarial fine-tuning steps. But according to papers that do second-party red-teaming, the SOTA is just a couple hundred steps.

To people working on adversarial vulnerabilities for safeguards against AI deepfake porn, I'm glad you're doing what you're doing. But don't forget that mitigations matter, & we're not always up against sophisticated attacks. Half the time, the perpetrators are literal teenagers.

🧵 Non-consensual AI deepfakes are out of control. But the 1st Amendment will likely prevent the US from directly prohibiting models/apps that make producing personalized NCII trivial. In this thread, I'll explain the problem and a 1st Amendment-compatible solution (I think).

One example of how easily harmful derivatives of open-weight models proliferate can be found on Hugging Face. Search "uncensored" or "abliterated" in the model search bar. You'll find some 7k models fine-tuned specifically to remove safeguards.

🚨 New paper from an awesome group led by Noam Kolt and @nickacaputo . We hear a lot about what important concepts and methods from AI research that lawyers need to understand. But it's really a two-way street... 🧵🧵🧵

Join us for our first CS seminar of the year, featuring @scasper.bsky.social! Learn more about his upcoming talk here: www.cs.jhu.edu/event/cs-sem... and check out other upcoming seminars here: www.cs.jhu.edu/department-s...

🧵Thanks in part to recent attention on Grok's widespread undressing, emerging consciousness around AI nudification apps is sparking discussions on making AI undressing apps illegal. Minnesota and the UK are currently actively considering laws that would do this.