Research

Multi-User Large Language Model Agents

Shu Yang², Shenzhe Zhu³, Hao Zhu¹, José Ramón Enríquez¹, Di Wang², Alex Pentland^1,4, Michiel A. Bakker⁴, Jiaxin Pei^1,†

¹Stanford University   ²KAUST   ³UT Austin   ⁴MIT   † Corresponding author

A systematic study of agents that serve several people at once, examining how current LLMs handle conflicting instructions, information boundaries, and coordination.

arXiv:2604.08567