🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

This week in The Algorithm, our weekly newsletter covering artificial intelligence: The EU wants to regulate your favorite AI tools. trib.al/nQf30ee

Absolutely delighted to finally be able share my new paper with Sandra Wachter [email protected] & Chris Russell "The Unfairness of fair machine learning: Levelling down and strict egalitarianism by default" papers.ssrn.com/sol3/papers.cf… Oxford Internet Institute Governance of Emerging Technologies ACM FAccT

I was part of the red team for GPT-4 — tasked with getting GPT-4 to do harmful things so that OpenAI could fix it before release. I've been advocating for red teaming for years & it's incredibly important. But I'm also increasingly concerned that it is far from sufficient. 🧵⤵️

Cho, Cho & Cho (2023) w/ @ejcho95 & Kyunghyun Cho finally online!😆 A short disclaimer for our work-in-progress: submitted to hal far before the recent decision on #FAccT2023 and was unexpectedly disclosed today after long moderation process,,, Next ver. should be much improved :-)

Thread: [1/4] Some MIT/Harvard collaborators and I just finished a project to show that Stable Diffusion objectively succeeds at copying the styles of digital artists with copyrighted work. Why might you care about this if you care about AI safety? github.com/thestephencasp…

Zuck @ Lex > Meta has solved human emulation > you can store and emulate your loved ones > Meta trying to figure out what they’ll allow for people who have passed > leaning into one creator -> many fans relationships through AI > the future is parasocial > they’re aiming

Very helpful analysis of the AI EO, from Karen Hao and Matteo Wong. "The White House’s Impossible AI Task". theatlantic.com/technology/arc…

About a week ago -- attended #EAAMO2023 and presented our PaperCard paper with @ejcho95 and Kyunghyun Cho Truely great people and community :) Thanks for all the constructive comments, discussions, and engaging interactions! arXiv version available in: arxiv.org/abs/2310.04824

My new nature paper "Role Play with Large Language Models" was briefly behind a paywall. It's now free-to-read: nature.com/articles/s4158…

What's the best work you know that finds evidence for internal representations of world models in a language model?

More hallucinations of case law by LLMs used in legal settings, this time in the UK. bailii.org/uk/cases/UKFTT…

In which I review two pieces I like lots: Risky Speech Systems: Tort Liability for AI-Generated Illegal Speech cyber.jotwell.com/risky-speech-s… Jotwell Jane Bambauer

The idea of "machine unlearning" is getting attention lately. Been thinking a lot about it recently and decided to write a long post: ai.stanford.edu/~kzliu/blog/un… 📰 Unlearning is no longer just about privacy and right-to-be-forgotten since foundation models. I hope to give a gentle

🚨 Fascinating paper alert: "Promises and pitfalls of artificial intelligence for legal applications" by Sayash Kapoor, Peter Henderson & Arvind Narayanan is a must-read for everyone interested in AI and the legal profession. Quotes: "Recent instruction-tuned language models (chatbots)

Excited to share a unifying formalism for the main problem I’ve tackled since starting my PhD! 🎉 Current AI Alignment techniques ignore the fact that human preferences/values can change. What would it take to account for this? 🤔 A thread 🧵⬇️

📢 New preprint: Aligning LLMs with diverse political views (with Philine Widmer Eunjung Cho Caglar Gulcehre and Elliott Ash) arxiv.org/pdf/2406.14155… More details in 🧵[1/7]

Our article “A Tough Balancing Act - The Evolving AI Governance in Korea”, where we provide an overview of the landscape of AI ethics and governance in Korea, has been published in East Asian Science, Technology and Society (EASTS). Freely available on: tandfonline.com/doi/full/10.10…

Will present this tomorrow, Nov. 14th 2pm at #EMNLP2024 (In-Person Poster Session G, Riverfront Hall). In case you're interested in generating diverse political viewpoints with LLMs, come to the poster session!

Okay, I'm starting a thread of all the best protest songs/covers I find in the run up to the impeachment vote, because Koreans know how to do protest songs and I think it could get epic. I'll add more as I find them. Let's start with this festive one: x.com/i/status/18648…

Really excited about our recent large collaboration work on NLP for Social Good. The work stems from our discussions at the NLP for Positive Impact Workshop (EMNLP 2024) Workshop at #EMNLP2024 EMNLP 2025. Thanks to all our awesome collaborators, workshop attendees and all supporters!