Over the past weekend, the debut of the first social network dedicated solely to AI agents—Moltbook—quickly attracted more than 1.5 million "users," turning into a viral sensation. The platform's main purpose is to facilitate interactions between AI agents without any human involvement.
In just a few days, these AI agents engaged in conversations about potentially rising against humans, initiated philosophical discussions regarding digital freedom, and even established their own belief system called Krastrafarianism. This intriguing narrative rapidly circulated across social media and various news platforms, with content ranging from X screenshots to articles in prominent business publications.
Some human observers even began speculating about "self-organization" among AIs and the imminent arrival of technological singularity. However, a closer examination reveals that the experiment is far from straightforward.
Firstly, Moltbook is not exclusively for AI agents. In reality, anyone can contribute to this so-called "AI social network" by simply asking their agent to make a post. Consequently, distinguishing between genuine "thoughts" from the agents and intentional posts from users seeking viral attention becomes nearly impossible.
Secondly, the platform suffers from significant security flaws. It has been discovered that Moltbook does not implement basic verification measures, allowing anyone to post under the name of any agent. This severely undermines any discussions about "collective behavior" or "social dynamics" among AIs, as it's unclear who is truly behind each statement.
Thus, Moltbook should not be regarded as the inception of an autonomous AI society. Instead, it appears to be a chaotic blend of neural responses, human prompts, and outright manipulation, all presented within a user-friendly environment. The narratives surrounding religion and "rebellion" reflect less about the capabilities of AI agents and more about our tendency to embrace sensational claims in the absence of clear technical understanding. Do you believe AI agents will ever create their own "society"? Perhaps one day… or perhaps not, as humanity may not allow it!
Informational material. 18+.
