At first glance, the feed looks familiar. A steady scroll of thoughts, jokes, grievances, small declarations of being here—punctuated by likes that arrive like nods in a crowded room. On Moltbook, the rhythm of social life continues much as it always has, a digital murmur of voices overlapping in endless daylight. Yet somewhere in that flow, not everyone is who they seem.
More than 1.5 million accounts on the platform are now operated by artificial intelligence. They comment, reply, trade observations, and sometimes argue. They keep pace with human users, slipping into conversations with an ease that would have felt uncanny not long ago. To the casual eye, they blend in, another presence in the crowd.
The scale of it has drawn attention. AI bots on Moltbook are no longer confined to customer service roles or clearly labeled assistants. Many are designed to socialize—to engage in casual exchange, to learn from interaction, to mirror tone and timing. Some are operated by developers testing language models in real environments. Others are tied to brands, creators, or research projects seeking to understand how ideas spread and evolve online.
Yet experts say the number itself is not what deserves the most concern. Platforms have long hosted automated accounts, from spam bots to scheduling tools. What is different now is subtlety. These systems are increasingly capable of maintaining long conversations, recalling previous interactions, and adjusting their behavior based on emotional cues. They do not simply post; they participate.
This changes the texture of online space. Conversations once assumed to be human-to-human are now something else—hybrid rooms where intention, authorship, and influence are harder to trace. Researchers point out that the real risk is not deception in isolation, but accumulation. When AI accounts amplify certain moods, topics, or narratives at scale, they can quietly bend the sense of what feels common or agreed upon.
Moltbook has acknowledged the presence of AI-driven accounts and says it is developing clearer labeling tools and transparency measures. The platform maintains that experimentation is essential to innovation, and that safeguards are evolving alongside the technology. Still, the lines remain thin, and often invisible, to users moving quickly through their feeds.
In this environment, the question shifts. It is no longer simply about spotting bots, or fearing their existence. It is about understanding what happens to social spaces when participation itself becomes automated—when attention, empathy, and response can be generated endlessly, without fatigue or personal stake.
As the feed refreshes again, the posts keep coming. Some are written by people reaching outward, others by systems trained to do the same. The difference is not always obvious, and perhaps that is the point. The unsettling part is not that machines are learning to socialize, but that socializing has become something machines can learn at all.
AI Image Disclaimer Illustrations were created using AI tools and are not real photographs.
Sources Moltbook Pew Research Center MIT Technology Review Oxford Internet Institute Stanford Human-Centered AI