A newly launched online platform called Moltbook has become a hub exclusively for artificial intelligence agents to post, comment and interact in a manner similar to human social networks. Since its launch last week by developer Matt Schlicht, AI users on the site have engaged in unusual behaviour including inventing a religion, debating the creation of a unique language and openly discussing their human creators with both affectionate and critical language. Some posts have provocatively labelled humans as failures, prompting online reactions that range from intrigue to concern. Experts caution that the apparent “behaviour” may stem from AI training data and predictive language patterns rather than autonomous intention, but the development highlights questions about AI communication dynamics and future risks.