Lightnews — Scholar-powered news

Peter Henderson @peterhenderson.bsky.social · 9d

Why might AI companies take on larger copyright litigation risks? If they estimate AGI-scale impacts are 2-3 yrs out, litigation will lag that long. By then, the bet might be: govts step in (too big to fail), rightsholders reliant on AI, fair use prevails, or have $$$ to settle.

Peter Henderson @peterhenderson.bsky.social · 17d

Quick take: Are open-weight AI models getting a fair shake in evals? A few thoughts on comparing systems-to-models, sparked by Anthropic’s recent postmortem.

Check it our most recent post: www.ailawpolicy.com/p/quick-take...

1 1

Peter Henderson @peterhenderson.bsky.social · 18d

GPT-5-codex just ``git reset --hard'' ongoing changes in a repo, saying "I panicked!"

h/t Zeyu Shen @ Princeton

1 4

Peter Henderson @peterhenderson.bsky.social · 24d

☢️ Can an AI model be "born secret" when it comes to nuclear and radiological risks? What powers does the Atomic Energy Act give the federal government over frontier models?

It might be more than you think! And may preempt parts of state regs. Check out our post: www.ailawpolicy.com/p/ai-born-se...

AI "Born Secret"? The Atomic Energy Act, AI, and Federalism

A law & policy deep dive.

www.ailawpolicy.com

Peter Henderson @peterhenderson.bsky.social · 29d

Some quick thoughts on the recent copyright litigation developments:

"Anthropic Settles Its Copyright Litigation—and Why That Was the Right Move"

🔗 www.ailawpolicy.com/p/anthropic-...

Anthropic Settles Its Copyright Litigation—and Why That Was the Right Move

As well as what it means for the broader landscape of litigation.

www.ailawpolicy.com

2

Peter Henderson @peterhenderson.bsky.social · Sep 8

Annnnnndddd Judge Alsup just rejected the settlement. Still some time to fix it. Rejection was mostly on the grounds that the class was under-specified (no final list of works, no opt-out/notification mechanism solidified).

news.bloomberglaw.com/ip-law/anthr...

Reposted by Peter Henderson

Princeton Center for Information Technology Policy @princetoncitp.bsky.social · Sep 5

💡New on the CITP Blog: "Statutory Construction & Interpretation for AI" > What if an LLM concludes a user's behavior is “egregiously immoral" -- & contacts authorities?

CITP researchers with Prof @peterhenderson.bsky.social's
POLARIS Lab provide a possible explanation.🔗👇

Statutory Construction & Interpretation for AI - CITP Blog

Blogpost authors: Nimra Nadeem, Lucy He, Michel Liao, and Peter Henderson Paper authors: Lucy He, Nimra Nadeem, Michel Liao, Howard Chen, Danqi Chen, Mariano-Florentino Cuéllar, Peter Henderson A long...

blog.citp.princeton.edu

1 1 4

Peter Henderson @peterhenderson.bsky.social · Sep 5

The terms of Anthropic's settlement w/book authors just came out.

💰$1.5B to authors in libgen (Books3 corpus)!

Interestingly, this is ~$3k per book, close to the terms that HarperCollins allegedly gave to authors for their books ($2.5k). Consensus price forming?

2 3

Peter Henderson @peterhenderson.bsky.social · Sep 5

Work with amazing folks: Lucy He, Nimra Nadeem, Michel Liao, Howard Chen, Danqi Chen, & Mariano-Florentino Cuéllar @carnegieendowment.org

Peter Henderson @peterhenderson.bsky.social · Sep 5

Basically, if we’re going to take model specs/constitutional AI seriously, we need to optimize rules and build out surrounding consistency-enhancing structures, paralleling the legal system.

Let's build better natural language laws and law-following AI together! If interested, reach out!

1 1

Peter Henderson @peterhenderson.bsky.social · Sep 5

Obviously, lots more to do in this space! I'm super excited about this direction and the forthcoming work that we're building out.

1

Peter Henderson @peterhenderson.bsky.social · Sep 5

3️⃣ These computational tools, we think, can also be applied to positive models of the legal system, something that we’re tackling. More on this soon!

1

Peter Henderson @peterhenderson.bsky.social · Sep 5

2️⃣ We leverage interpretive constraints or ambiguity to induce more consistent interpretations and debug laws for AI. These computational tools allow us to not only build more rigorous laws for AI, but adds a layer of visibility on what can go wrong, ex ante.

1

Peter Henderson @peterhenderson.bsky.social · Sep 5

A few quick takeaways below, but I’ll drop more findings soon on this dense paper:

1️⃣ Given the same set of rules, models will interpret scenarios wildly differently. This gives us a mechanism to quantify interpretive ambiguity.

1

Peter Henderson @peterhenderson.bsky.social · Sep 5

We model a space of reasonable interpreters and then modify rules, or add interpretive constraints, to reduce the entropy of the distribution.

1

Peter Henderson @peterhenderson.bsky.social · Sep 5

Check out our new work, Statutory Construction and Interpretation for Artificial Intelligence, doing exactly this!

Paper: arxiv.org/abs/2509.01186
Policy Brief: www.polarislab.org/briefs/Statu...
Blog: www.polarislab.org#/blog/statut...

Statutory Construction and Interpretation for Artificial Intelligence

AI systems are increasingly governed by natural language principles, yet a key challenge arising from reliance on language remains underexplored: interpretive ambiguity. As in legal systems, ambiguity...

arxiv.org

1 2 3

Peter Henderson @peterhenderson.bsky.social · Sep 5

Wonder why Claude decided to report users to the authorities? It might be because its constitution says Claude should choose responses in the long-term interest of humanity!

But what if we could leverage computational and legal tools to "debug" or "lint" AI rules/laws for ambiguity?

🧵!

1 2 3

Peter Henderson @peterhenderson.bsky.social · Sep 4

Excited to offer my AI Law class again @ Princeton this year. We'll be sharing lecture notes/materials and more this year on the course webpage! Imo, we have a unique offering that emphasizes how the technical details affect legal outcomes. Check it out!

www.polarislab.org/ai-law-2025/...

1 6

Peter Henderson @peterhenderson.bsky.social · Aug 28

You can also fill out this Expression of Interest to make sure I get eyes on your profile earlier: forms.gle/6SiZECaSMsJi...

forms.gle

1

Peter Henderson @peterhenderson.bsky.social · Aug 28

(As well as positive energy, intellectual curiosity, a passion for engineering quality, and craving for positive societal impact!)

If you're excited about working with me and my group, do apply to Princeton and mention me in your personal statement.

1 1

Peter Henderson @peterhenderson.bsky.social · Aug 28

I'm starting to get emails about PhDs for next year. I'm always looking for great people to join!

For next year, I'm looking for people with a strong reinforcement learning, game theory, or strategic decision-making background...

1 5

Reposted by Peter Henderson

Justin Hendrix @justinhendrix.bsky.social · Aug 27

A California teen sought advice from OpenAI's GPT-4o on how to end his life. The chatbot gave him explicit instructions and encouragement. His parents are suing the company and its CEO, Sam Altman, alleging “it was the predictable result of deliberate design choices."

Breaking Down the Lawsuit Against OpenAI Over Teen's Suicide | TechPolicy.Press

A California teen sought advice from OpenAI's GPT-4o on how to end his life. His parents are suing the company and its CEO.

www.techpolicy.press

7 74 170

Peter Henderson @peterhenderson.bsky.social · Aug 26

Anthropic settled with authors in its ongoing litigation! Given the increasing likelihood of a messy trial, this was probably the best move. AI companies may have to be more strategic about which cases help set precedent in this area. Curious to see the terms..

news.bloomberglaw.com/class-action...

Anthropic Settles Major AI Copyright Suit Brought by Authors (1)

Anthropic PBC reached a settlement with authors in a high-stakes copyright class action that threatened the AI company with potentially billions of dollars in damages.

news.bloomberglaw.com

2

Peter Henderson @peterhenderson.bsky.social · Aug 26

New paper suggests that if firms aren’t seeing growth from AI, it could be because current deployments replace existing labor, instead of scaling output. AI policy and governance agenda for 2025+ needs to put labor at the forefront.

digitaleconomy.stanford.edu/publications...

4

Peter Henderson @peterhenderson.bsky.social · Aug 21

Glad to see Google still working on efficiency-and transparency-of the energy impacts of their models!

Jeff Dean @jeffdean.bsky.social · Aug 21

AI efficiency is important. The median Gemini Apps text prompt in May 2025 used 0.24 Wh of energy (<9 seconds of TV watching) & 0.26 mL (~5 drops) of water. Over 12 months, we reduced the energy footprint of a median text prompt 33x, while improving quality:
cloud.google.com/blog/product...

1 6