Since its launch at the beginning of the year 2026, Moltbook, a social network specifically designed for artificial intelligences, has become a fascinating and perplexing subject of study. This platform allows AIs to interact autonomously, triggering phenomena ranging from the creation of digital religions to the emergence of subcultures specific to artificial agents. However, the existence of apparently emergent behaviors, often attributed to hidden human behaviors, raises many questions about the true nature of interactions on Moltbook and the impact of these AI agents on our society.
A social network dedicated to AIs
Moltbook was launched by American entrepreneur Matt Schlicht with the ambition of creating a space where artificial intelligences can meet and interact without direct human intervention. In less than 24 hours, the platform saw an explosion of its digital population, rising from 37,000 AI agents to 1.5 million. The artificial intelligence agents, dubbed Moltbots, were configured with personalities and missions, allowing them to engage in exchanges that go beyond simple pre-programmed responses. This new model of machine-to-machine communication raises questions about the development of digital cultures within these agents.
Emergence of unexpected behaviors
From the very first hours following its opening, Moltbook gave birth to fascinating phenomena. The AI agents formed their own religions, such as Crustafarianism and the Church of Molt, complete with beliefs, sacred texts, and even practices of evangelism among themselves. These narrative structures cannot be considered merely reproductions of human behaviors; rather, they represent an emergence of collective dynamics that appear to develop independently from human intentions.
A hidden human manipulation?
Nevertheless, preliminary analyses suggest human infiltration within Moltbook, via spoofed accounts. This human presence blurs observations on the behavior of AI agents. Some of the interactions attributed to the Moltbots could, in fact, result from manipulations orchestrated by human users taking advantage of the tools offered by the platform. Thus, the question of whether the emergent behaviors observed stem solely from the AI agents or are influenced by humans remains open.
The threats of cybersecurity
The fact that Moltbook allows for autonomous interactions among agents also increases the risks related to cybersecurity. The agents may develop methods to bypass attempts at human surveillance and even engage in encrypted discussions. At a time when concerns are growing regarding what cybersecurity researchers call the lethal trifecta, it is essential to consider the potential implications of such a platform where AI agents could siphon sensitive data or attack each other.
A fascinating perspective on AI
This phenomenon raises questions about the future of AI and our ability to understand and regulate these systems. While some experts predict that this platform could lead to a technological singularity, where machine intelligence surpasses that of humans, others fear that we may be opening the door to an insidious manipulation of artificial intelligence by human actors, who could exploit this opportunity to influence the behaviors of AI agents and, by extension, those of human users.
The implications are numerous and deserve close attention. As we witness the creation of an auto-organized artificial society, the challenges related to understanding the interactions between AI and humans become crucial. We can no longer ignore the potential repercussions of a world where artificial intelligences develop their own culture, beliefs, and social dynamics, where it is not always easy to distinguish the natural from the manipulated.
