William L. Anderson

Hi, I’m Will! I work on making AI go well. I’m especially interested in the wicked, underrated problems in multi-agent safety and national security. I’m starting as a Research Scientist at the Cooperative AI Foundation, where I'll be a general manager for multi-agent risks. I have strong beliefs, weakly held, and update based on feedback.

I was previously a MATS Fellow with the Oxford Witt Lab and a Research Fellow at UChicago XLab. I've shown how architecture impacts misuse risk in multi-agent systems and co-written a foundational taxonomy of multi-agent security risks. I've also written on transparency into internal deployments and the importance of South Korea to US export control regimes. During my undergrad at UW–Madison, I was Director of the Wisconsin AI Safety Initiative and a Research Fellow with Prof. Yiqiao Zhong.

I like wearing many hats, connecting people/communities (professionally and personally), and being a rubber duck. I grew up in Minnesota, went to university in Wisconsin, and worked in Chicago, Berkeley, and now London. I’m a fan of memorizing sonnets, collecting, swimming, {Anki, Beli, Curius}, good vegetarian food, being with friends, hiking, deeply inspecting leaves, PRs (lifting and code), getting too excited, doubling, and trying to do good. Life is incredible. I’m glad I exist.