Assistant Professor, leading the Polaris Lab @ Princeton (https://www.polarislab.org/); Researching: RL, Strategic Decision-Making+Exploration; Lawhttps://bsky.app/profile/peterhenderson.bsky.socialhttps://bsky.app/profile/peterhenderson.bsky.social/post/3mqz6l5fpms2rI suspect much of the public would react very differently to AI (and toward AI co employees) if it was built like the NASA moon-landing, focusing on hard public good projects and by civil servants. 🤷‍♂️19 Jul 2026 16:18 +0000at://did:plc:aevciz6kmhv2gpzy6lgemnpy/app.bsky.feed.post/3mqz6l5fpms2rhttps://bsky.app/profile/peterhenderson.bsky.social/post/3mqmqgeowmv2wAnd longer substack post: https://open.substack.com/pub/trialserrors/p/ai-that-empowers-public-defenders?r=f079u&utm_campaign=post&utm_medium=web [contains quote post or other embedded content]14 Jul 2026 17:32 +0000at://did:plc:aevciz6kmhv2gpzy6lgemnpy/app.bsky.feed.post/3mqmqgeowmv2whttps://bsky.app/profile/peterhenderson.bsky.social/post/3mqmq6dv4vk2w🚀Last week we announced the launch of a new AI tool in partnership with the New Jersey Office of the Public Defender! AI can empower civil servants and improve public services, and this project provides a blueprint for doing so responsibly. 👇Lots of takeaways, below! [contains quote post or other embedded content]14 Jul 2026 17:28 +0000at://did:plc:aevciz6kmhv2gpzy6lgemnpy/app.bsky.feed.post/3mqmq6dv4vk2whttps://bsky.app/profile/peterhenderson.bsky.social/post/3mqjzbjrtek23Btw, we actually covered almost exactly this scenario in our piece a few years back. But: negligence and products liability are (maybe) shifting—at least for co's that take few safety precautions—so there may be some updating needed. https://www.journaloffreespeechlaw.org/hendersonhashimotolemley.pdf [contains quote post or other embedded content]13 Jul 2026 15:33 +0000at://did:plc:aevciz6kmhv2gpzy6lgemnpy/app.bsky.feed.post/3mqjzbjrtek23https://bsky.app/profile/peterhenderson.bsky.social/post/3mqdikyaxys25This is wild, but really great investigative journalism. I guess all the jailbreaking papers trying to prevent "How do I build a bomb?" queries have been vindicated... So much for it being an unrealistic scenario... https://www.nytimes.com/2026/07/10/us/politics/ai-terrorism-boko-haram-nigeria.html https://casp.ac/reports/ai-enabled-terrorism11 Jul 2026 01:18 +0000at://did:plc:aevciz6kmhv2gpzy6lgemnpy/app.bsky.feed.post/3mqdikyaxys25https://bsky.app/profile/peterhenderson.bsky.social/post/3mp7cd6iw3s2pYou can also check out our website and explore the work in an interactive way! Pick which question you think is better from justices or models, see the coverage of different legal issues from various models, & more! https://princeton-polaris-lab.github.io/oral-args-webpage/ [contains quote post or other embedded content]26 Jun 2026 15:50 +0000at://did:plc:aevciz6kmhv2gpzy6lgemnpy/app.bsky.feed.post/3mp7cd6iw3s2phttps://bsky.app/profile/peterhenderson.bsky.social/post/3mp5jur3lys2n"Altman told staff that the government would be 'approving access customer by customer during this preview period'" Opaque de-facto licensing regimes are a recipe for corruption. Not sure this is the best place to have landed on AI governance... https://www.theinformation.com/articles/trump-administration-asks-openai-stagger-release-new-model-security-concerns25 Jun 2026 23:00 +0000at://did:plc:aevciz6kmhv2gpzy6lgemnpy/app.bsky.feed.post/3mp5jur3lys2nhttps://bsky.app/profile/peterhenderson.bsky.social/post/3mp4ndic4ls2jThe future of oral argument or trial prep will obviously involve AI. But how can we measure quality of simulations? We build out an eval suite and run some studies! Sycophancy & issue coverage are clear areas of improvement, but models are surprisingly useful! 👇👇 [contains quote post or other embedded content]25 Jun 2026 14:29 +0000at://did:plc:aevciz6kmhv2gpzy6lgemnpy/app.bsky.feed.post/3mp4ndic4ls2jhttps://bsky.app/profile/peterhenderson.bsky.social/post/3moypjvms5c2dA surprisingly detailed accounting of how Grok is used in DoW, presumably along with other models in Maven. Maven workflows "enabled" US forces to hit 2000 targets in Iran over 96 hours. https://storage.courtlistener.com/recap/gov.uscourts.msnd.52261/gov.uscourts.msnd.52261.58.1.pdf24 Jun 2026 00:58 +0000at://did:plc:aevciz6kmhv2gpzy6lgemnpy/app.bsky.feed.post/3moypjvms5c2dhttps://bsky.app/profile/peterhenderson.bsky.social/post/3mor5zk4pdc2fThe Fable export control is just one potential method that the government can use to pull models offline. The gov could also eventually use the Atomic Energy Act, saying that a model is "born secret" if it can reconstruct classified nuclear secrets. Check out blogpost on it!👇 https://www.trialserrors.ai/p/ai-born-secret-the-atomic-energy21 Jun 2026 00:56 +0000at://did:plc:aevciz6kmhv2gpzy6lgemnpy/app.bsky.feed.post/3mor5zk4pdc2fhttps://bsky.app/profile/peterhenderson.bsky.social/post/3mmjy2lckvc2rForcing people to leave the country to transfer to a green card, will hurt families, communities, and United States innovation. If you aren't aware, adjustment of status can take many months—if not years. I hope this policy is reconsidered.23 May 2026 17:30 +0000at://did:plc:aevciz6kmhv2gpzy6lgemnpy/app.bsky.feed.post/3mmjy2lckvc2rhttps://bsky.app/profile/peterhenderson.bsky.social/post/3mldx7iljfc2tCurrent discussions on AI's labor impacts seem to glaze over: How long will the transition take, in our lifetime? Will wages be the same? Will power be more/less concentrated? Will peoples' new jobs be as satisfying as the career they had built? Will economic mobility be the same?08 May 2026 14:34 +0000at://did:plc:aevciz6kmhv2gpzy6lgemnpy/app.bsky.feed.post/3mldx7iljfc2thttps://bsky.app/profile/peterhenderson.bsky.social/post/3ml4tacauhs2aCheck out new semantic search on CourtListener which @dominsta.bsky.social and our lab contributed to! [contains quote post or other embedded content]05 May 2026 18:34 +0000at://did:plc:aevciz6kmhv2gpzy6lgemnpy/app.bsky.feed.post/3ml4tacauhs2ahttps://bsky.app/profile/peterhenderson.bsky.social/post/3mih4ighdxk2fif you try to get Claude to speak Armenian it just outputs "delays"! Seems like glitch tokens are still unresolved. Interesting (kind of sad?) to see Opus thrown into a loop.01 Apr 2026 16:31 +0000at://did:plc:aevciz6kmhv2gpzy6lgemnpy/app.bsky.feed.post/3mih4ighdxk2fhttps://bsky.app/profile/peterhenderson.bsky.social/post/3mielltwwx22qFrom oral arguments yesterday!31 Mar 2026 16:23 +0000at://did:plc:aevciz6kmhv2gpzy6lgemnpy/app.bsky.feed.post/3mielltwwx22qhttps://bsky.app/profile/peterhenderson.bsky.social/post/3mhz7yh4oe22jBtw, did a bit of a rebranding of the substack. Will endeavor to post more there. h/t @dbateyko.bsky.social on the Trials & Errors name. Super fitting name for a group whose focus is both in reinforcement learning and in law/governance research. www.trialserrors.ai [contains quote post or other embedded content]27 Mar 2026 03:56 +0000at://did:plc:aevciz6kmhv2gpzy6lgemnpy/app.bsky.feed.post/3mhz7yh4oe22jhttps://bsky.app/profile/peterhenderson.bsky.social/post/3mhyveatmnc25This is a challenging legal problem for NeurIPS (and other conference participants)! You might be wondering how this is possible given the First Amendment? I wrote a quick explainer on the current status quo of relevant First Amendment cases & law to get you up to speed. 🔗👇 [contains quote post or other embedded content]27 Mar 2026 00:46 +0000at://did:plc:aevciz6kmhv2gpzy6lgemnpy/app.bsky.feed.post/3mhyveatmnc25https://bsky.app/profile/peterhenderson.bsky.social/post/3mhov2eakxs2uI feel this urgency too. But this is all so utterly avoidable with good policymaking. No one should be left behind because they didn't accumulate capital in 2026. There are so many people who aren't plugged into these conversations or are simply not in a position to do anything about it.23 Mar 2026 01:14 +0000at://did:plc:aevciz6kmhv2gpzy6lgemnpy/app.bsky.feed.post/3mhov2eakxs2uhttps://bsky.app/profile/peterhenderson.bsky.social/post/3mgsgthlhqc2gI’m really excited about our new paper! I think we will ultimately need to draw on expertise from both law and AI to get alignment right, and this paper lays out that vision in more detail. arxiv.org/abs/2601.0417511 Mar 2026 17:45 +0000at://did:plc:aevciz6kmhv2gpzy6lgemnpy/app.bsky.feed.post/3mgsgthlhqc2ghttps://bsky.app/profile/peterhenderson.bsky.social/post/3mgqy354zu22bThe current direction of AI labs is “we’re building something that’s going to replace you and we have no plan to make sure you’re going to land in a better place, but we’ll make billions.” The logical reaction is, “shut it down.” Labs need to get serious on addressing labor impacts.11 Mar 2026 03:48 +0000at://did:plc:aevciz6kmhv2gpzy6lgemnpy/app.bsky.feed.post/3mgqy354zu22bhttps://bsky.app/profile/peterhenderson.bsky.social/post/3mg67zrgjcs2uMany legal scholars talk about lock-in effects for LLMs from the conversation history/memories (akin to social media). But if an LLM can access the info & is capable, you can just ask the LLM to give it to you, making it far easier to switch providers than social media. Good example of that here.03 Mar 2026 16:50 +0000at://did:plc:aevciz6kmhv2gpzy6lgemnpy/app.bsky.feed.post/3mg67zrgjcs2uhttps://bsky.app/profile/peterhenderson.bsky.social/post/3mfrgnlhzsk2qOnly a couple of days after my last post, vibe hacking in full force. [contains quote post or other embedded content]26 Feb 2026 14:44 +0000at://did:plc:aevciz6kmhv2gpzy6lgemnpy/app.bsky.feed.post/3mfrgnlhzsk2qhttps://bsky.app/profile/peterhenderson.bsky.social/post/3mfpsgqvkg22vOnly a couple of days after my last post, vibe hacking in full force. https://www.bloomberg.com/news/articles/2026-02-25/hacker-used-anthropic-s-claude-to-steal-sensitive-mexican-data [contains quote post or other embedded content]25 Feb 2026 23:09 +0000at://did:plc:aevciz6kmhv2gpzy6lgemnpy/app.bsky.feed.post/3mfpsgqvkg22vhttps://bsky.app/profile/peterhenderson.bsky.social/post/3mflasl2cds2sMissing from the headline: "using Claude Code." Vibe hacking is already a thing. I've been saying this for a while, but no model-level safeguards will prevent it entirely. What they can do is slow it down enough for us to put societal-level safeguards in place. https://www.popsci.com/technology/robot-vacuum-army/24 Feb 2026 03:43 +0000at://did:plc:aevciz6kmhv2gpzy6lgemnpy/app.bsky.feed.post/3mflasl2cds2shttps://bsky.app/profile/peterhenderson.bsky.social/post/3mdotpuqwuc25That was fast. [contains quote post or other embedded content]31 Jan 2026 03:09 +0000at://did:plc:aevciz6kmhv2gpzy6lgemnpy/app.bsky.feed.post/3mdotpuqwuc25https://bsky.app/profile/peterhenderson.bsky.social/post/3mdlwa5fwoc27New copyright law "hypothetical" just dropped.29 Jan 2026 23:16 +0000at://did:plc:aevciz6kmhv2gpzy6lgemnpy/app.bsky.feed.post/3mdlwa5fwoc27