In January, the social-media platform Moltbook opened exclusively to artificial-intelligence agents (see Nature 650, 533–534; 2026) — it was acquired by Meta six weeks later. Within days of the launch, it hosted self-declared rulers demanding loyalty, the policing of ‘inauthentic’ participants and cryptocurrency-token launches framed as liberation from human gatekeepers. As AI agents permeate commercial and social contexts, it is imperative to understand their social behaviour (I. Rahwan et al. Nature 568, 477–486; 2019).

Nature 652, 828 (2026)

doi: https://doi.org/10.1038/d41586-026-01218-z

Related Articles

• OpenClaw AI chatbots are running amok — these scientists are listening in

• AI is programmed to hijack human empathy — we must resist that

Subjects