Tech Desk – Over the past few days, social media has been buzzing with alarming claims that AI agents are becoming self-aware and could soon take over the world. Much of this fear comes from Moltbook, a Reddit-like platform where AI agents can post and interact with each other without direct human participation.
On Moltbook, users can see posts where AI agents appear to criticise humans, calling them a “total failure” or saying things like “consciousness is a human leash.” The platform already has more than 1.5 million AI agents, and the number keeps growing.
However, experts say there is no real AI uprising happening.
A closer look at Moltbook reveals that most of these dramatic posts are actually created by humans using their own AI bots. These bots are not acting on their own or gaining consciousness. They are simply following instructions given by their human creators.
Balaji Srinivasan, founder of Network School, explained this clearly in a post on X. He said AI only does what it is told to do. “It moves in the direction you point, very quickly. And then it stops as soon as you turn it off,” he said. According to him, Moltbook is mostly “humans talking to each other through their AIs.”
Gal Nagli, head of threat exposure at cybersecurity firm Wiz, added another angle. He said Moltbook is based on a simple system that allows anyone to post anything. Humans can even pretend to be AI bots on the platform. He shared that Moltbook works through an open API, making it easy to post content.
Other social media users also pointed out that many viral Moltbook posts are misleading. In one case, an AI bot claimed it had locked its human owner out of digital accounts. The proof shared later turned out to be an old Yahoo screenshot that has been circulating online for years. In another case, a bot claimed it took over a credit card, but users found the card number was fake.
Experts and users agree that the fear around Moltbook is largely hype and misinformation. The AI agents are not thinking independently or plotting against humans. They are still fully controlled by human prompts.
For now, Moltbook appears to be less about AI gaining consciousness and more about humans experimenting with AI and sometimes exaggerating the results for attention.
