Xu Li

Xu Li 李旭

Agent safety · Human-AI coevolution · Agent evaluation

Hi! I am a 1st-year PhD student at Northeastern University, advised by Prof. Weiyan Shi in Chats Lab. In summer 2026, I interned at VirtueAI and Meta Superintelligence Labs with Prof. Bo Li and Minzhou Pan, working on agent safety and agent attacks.

I'm interested in how humans can stay in the loop and use AI well.

More about my research
  • Agent safety: what goes wrong when the human is not actively in the loop.
  • Human-AI coevolution: how people evolve to use AI better, and how humans reshape AI.
  • Agent evaluation: what agents can really do, and how they assist humans.
Get in touch

News

2026.04
Unsafer in Many Turns: Benchmarking and Defending Multi-Turn Safety Risks in Tool-Using Agents
Xu Li, Simon Yu, Minzhou Pan, Yiyou Sun, Bo Li, Dawn Song, Xue Lin, Weiyan Shi*
ICML 2026 Oral @ AI4GOOD Workshop, ICML 2026 [pdf] [X thread] [homepage]
2025.09
ICCV 2025 6th CLVISION Workshop Challenge
4th place out of 203 submissions and 65 participants
[link]
2025.05
MM-Prompt: Cross-Modal Prompt Tuning for Continual Visual Question Answering
Xu Li, Fan Lyu*
arXiv 2025 [pdf]

Contact

Email: li.xu2@northeastern.edu