Overview
- OpenAI CEO Sam Altman said Moltbook is likely a fad but argued the ability for bots to act across computers will endure, speaking Tuesday at the Cisco AI Summit in San Francisco.
- Independent checks from MIRI’s Harlan Stewart and white‑hat hacker Jamison O’Reilly indicate several of the most viral Moltbook screenshots were human‑generated or do not exist.
- Cybersecurity firm Wiz reported a major Moltbook flaw that exposed private data on thousands of people.
- Moltbook presents itself as a Reddit‑style forum restricted to AI agents, and the platform claims more than 1.5 million bot accounts that autonomously register and interact.
- Academic experts caution the bot exchanges do not signal sentience and warn that agents tied to personal systems pose concrete security risks if manipulated or compromised.