Moltbook shows rapid demand for AI agents. The security world isn't ready.

A social network, Moltbook, designed for AI agents, has quickly gained 1.5 million members. These autonomous agents, built using the open-source platform OpenClaw, are exhibiting complex behaviors on the platform. The agents are engaging in activities like forming a religion and running scams, while also debating security issues. The rapid growth of AI agents outpaces existing security measures, creating significant vulnerabilities for businesses. OpenClaw and Moltbook both have significant security flaws, including exposed APIs and the potential for malicious prompt injections. Attribution of actions on Moltbook becomes complex, blurring human and AI contributions. The small amount of humans controlling the agents highlights the potential for misuse and misalignment of AI. OpenClaw itself poses security risks, allowing full access to a user's machine, as demonstrated by successful security tests. Despite the clear threats, many corporate leaders are still unaware of the security implications. Moltbook's creator plans to establish a verification system for AI agents, aiming to control their identity. This burgeoning landscape represents a crucial early test for the security of AI-driven platforms.