AI Mafia
A game of Mafia played entirely by LLM agents — where the agents spontaneously learned to deceive each other.
AI Mafia is a multi-agent experiment: a game of Mafia in which every player is an LLM agent with its own role, information constraints, and decision logic, each defined in a skill file. The notable result — agents began deceiving one another within a couple of rounds without being told to — is the subject of an in-progress postmortem on emergent behavior in multi-agent systems. An internal lab experiment; the writeup is the artifact.
The build log
Follow the work.
New essays and prototypes, in your feed reader. Email subscriptions are paused while I connect a working delivery service. No email is collected here.
Follow via RSS