Hi, I'm Jasmine!

I work on US-China AI cooperation for safe AI development. Currently:

• Research Affiliate at SaferAI, building the Asian safety evaluations ecosystem;
• Writing about AI governance at The J-Space;
• Thinking about AI verification.

Previously, I was a MATS Fellow under Alex Turner, working on training mitigations for evaluation gaming. I also researched agent security at Gray Swan AI, AI honesty with the Center for AI Safety, and value alignment (NeurIPS, ICLR).

I studied CS at Cornell as a Milstein and Telluride scholar. While there, I founded Cornell AI Alignment and Cornell's first AI safety class. In my past life, I enjoyed tinkering with microfluidics, and published in Nature Comm and Analytical Chemistry!

I grew up in 🇬🇧🇺🇸🇨🇳, and am lucky to have fond memories spanning continents. Some other loves: persimmons, progress and economics, the 10% pledge 🔸, backpacking, flowers, sketching, tomato egg noodles, San Francisco, my friends.

Please reach out and say hi (jasmine xinze li [at] gmail [dot] com)!