OpenClaw is a major leap forward for AI—and a cybersecurity nightmare

Cybersecurity researchers have discovered roughly 1,000 unprotected gateways to OpenClaw, an open-source and proactive AI agent that can be controlled through text conversations with apps like WhatsApp or Telegram. The gateways were found on the open internet, allowing anyone to access users’ personal information. One white hat hacker also reportedly gamed OpenClaw’s skills system, which lets users add plugins for tasks like web automation or system control, to reach the top of the rankings and be downloaded by users around the world.