Peter Henderson

🚀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!

This work would not be possible without @dominsta_nlp , in partnership with NJOPD, the NJ Innovation Authority, the NJ Office of Information Technology, AWS engineers, and public defenders who shared their expertise.

🚀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!

These new evals also showed that most open-weight embedding models fall short of being useful in this context... so we ended up fine-tuning models using synthetic data to shore them up for this real-world use case!

🚀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!

Before doing this we conducted structured interviews to identify concerns. One identified risk related to hallucinations. As a result, we focused on an extractive-first system, with optional generative summaries for only the facts of the case to situate the retrieved content.

🚀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!

Public defenders manage large caseloads and tight deadlines. They also rely on a wealth of shared knowledge: past briefs, memos, directives, and lessons from colleagues about what has—and hasn’t—worked in court. Here's an example of an AG directive memo.

The 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!👇

Link: www.trialserrors.ai/p/ai-born-se...

Current 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?

if 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.

From oral arguments yesterday!

This 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. 🔗👇

I 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.

Single parents working three jobs to make ends meet cannot possibly work harder to accumulate capital. They already work hard enough as it is. People in this position should not be "left behind." There should be no "permanent underclass,” as many are worried about.

I’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/2601.04175

Many 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.

Missing 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. www.popsci.com/technology/r...

As models discover more exploits, we may need something like a responsible disclosure period for major jumps in cyber capabilities. Before release, trusted defenders get privileged access to the more capable model. Together, they scan for vulnerabilities at scale & notify as many as possible.