A. Feder Cooper

I’ll be speaking on the panel, then presenting the poster with several of my co-authors.

Joint w/ @marklemley.bsky.social @mbogen.bsky.social @nicolaspapernot.bsky.social @klyman.bsky.social @milesbrundage.bsky.social @jtlg.bsky.social @hannawallach.bsky.social @zephoria.bsky.social + many others

Co-led with @katherinelee.bsky.social, this cross-institution, cross-disciplinary collaboration details the challenges of using machine unlearning in generative AI to meet substantive goals in law and policy (with a particular focus on copyright, privacy, and safety contexts).

[NeurIPS '25] Our oral slot and poster session on "Machine Unlearning Doesn't Do What You Think: Lessons for Generative AI Policy and Research" are tomorrow, December 4! [https://arxiv.org/abs/2412.06966]

Oral: 3:30-4pm PST, Upper Level Ballroom 20AB

Poster 1307: 4:30:-7:30pm PST, Exhibit Hall C-E

Tutorial tomorrow at 1:30PM PST!

My talk slots will cover memorization + copying in models and their outputs, canonical extraction methods, and recent work with @marklemley.bsky.social and others on extracting pieces of memorized books from open-weight models.

arxiv.org/abs/2505.12546

I'm at NeurIPS & hiring for our pretraining safety team at OpenAI! Email me if you want to chat about making safer base models!

Comparison requires valid measurement: Rethinking attack success rate comparisons in AI red teaming

Poster #1110 | Fri Dec 5, 4:30-7:30pm PST, Exhibit Hall C,D,E

openreview.net/forum?id=d7h...

Machine Unlearning Doesn't Do What You Think: Lessons for Generative AI Policy and Research "Machine unlearning" is a popular proposed solution for mitigating the existence of content in an AI model that is problematic for legal or moral reasons, including privacy, copyright, safety, and mor...

Machine Unlearning Doesn't Do What You Think: Lessons for Generative AI Policy and Research

Oral | Thu Dec 4 3:30-4pm PST, Upper Level Ballroom 20AB

Poster #1307 | Thu Dec 4, 4:30:-7:30pm PST, Exhibit Hall C,D,E

arxiv.org/abs/2412.06966

The Common Pile v0.1: An 8TB Dataset of Public Domain and Openly Licensed Text Large language models (LLMs) are typically trained on enormous quantities of unlicensed text, a practice that has led to scrutiny due to possible intellectual property infringement and ethical concern...

The Common Pile v0.1: An 8TB Dataset of Public Domain and Openly Licensed Text

Poster #102 | Fri Dec 5, 11am-2pm PST, Exhibit Hall C,D,E

arxiv.org/abs/2506.05209

Exploring the limits of strong membership inference attacks on large language models State-of-the-art membership inference attacks (MIAs) typically require training many reference models, making it difficult to scale these attacks to large pre-trained language models (LLMs). As a resu...

Exploring the limits of strong membership inference attacks on large language models

Poster #1300 | Fri Dec 5, 11am-2pm PST, Exhibit Hall C,D,E

arxiv.org/abs/2505.18773

Tutorial: Data Privacy, Memorization, & Legal Implications in Generative AI

Tue Dec 2, 1:30-4pm PST, Exhibit Hall F

w/ @pratyushmaini.bsky.social + Joe Gratz

memlaw-tutorial.github.io

Excited to be at NeurIPS this week in San Diego! Please reach out (best over email) if you’d like to chat about privacy & security, scalable evals, and reliable ML systems.

I’ll be presenting a few papers/speaking at some events, please stop by! Will post details throughout the week (summary below)

📣 Postdocs at Yale FDS! 📣 Tremendous freedom to work on data science problems with faculty across campus, multi-year, great salary. Deadline 12/15. Spread the word! Application: academicjobsonline.org/ajo/jobs/31114 More about Yale FDS: fds.yale.edu

Just finished reading the GEMA v. OpenAI decision (slowly, my German isn't great). Looks like a not small part of the analysis tracked parts of arguments @jtlg.bsky.social and I made in 2024.

I don't have a well-formed response yet, but hopefully soon. (Main thought atm is a very unpolished "woah")

Today's decision in GEMA v. OpenAI by a German court holds that ChatGPT infringes copyright when it memorizes song lyrics. The opinion cites my paper with @afedercooper.bsky.social on memorization in generative models, and its analysis tracks ours.

drive.google.com/file/d/1dUaD...

Omg I cackled

Bill Ackman gotta be on the third draft of a tweet longer than Middlemarch right now

happy to dm you about it :)

I could see why someone not actively doing research in that subfield would very reasonably think that. But as someone who used to publish there, I’ll just say “lol”

I’m kinda known as a copyright person, but (even in memorization) I mainly study how to draw reliable conclusions from large-scale AI/ML systems. There’s a long spiel why, but today I feel defeated. 100 hours/week on this for 6 years, just to find out a parent treats Gemini in search as ground-truth

what!! scholar counts the same paper more than once! that's insane! I guess i've noticed this and merged them, e.g., if I've changed the title in revisions. but i assumed once you did that they de-duped any citations. because you're telling them it's the same paper...

lol

what irony that I often feel like the imposter, producing only 6-8 papers as a core contributor per year

I just feel particular dumb not to have considered this, seeing as I do a lot of work on manipulating ML models. This just would never have occurred to me as a thing that was worth doing.
This is just...so (likely usually) petty and stupid.

Omg what

✨✨

This is a really great community of researchers, and every accepted paper gets a generously long talk slot to present.

CFP: computersciencelaw.org/2026-2/2026-...

Main track deadline (archival and non-archival): September 30, AoE

The NeurIPS position track didn't take a large number of extraordinary papers that surpassed the acceptance bar, limiting the acceptance rate to an unusually low 6%.

If you have a rejected paper at the intersection of ML and law, consider submitting to ACM CSLaw '26.

Machine Unlearning Doesn't Do What You Think: Lessons for Generative AI Policy, Research, and Practice We articulate fundamental mismatches between technical methods for machine unlearning in Generative AI, and documented aspirations for broader impact that these methods could have for law and policy. ...

Our paper "Machine Unlearning Doesn't Do What You Think" was accepted for presentation at NeurIPS

Congrats @afedercooper.bsky.social and @katherinelee.bsky.social, who led the effort

arxiv.org/abs/2412.06966

One more week to submit to CSLaw '26!!

at least 100k with all the appendices 🙃

AI Copyright Lawsuits with Pam Samuelson | Scaling Laws

For an update on the state of play in the generative AI copyright cases, try this podcast: shows.acast.com/arbiters-of-...