I loved my OpenClaw AI agent – until he turned against me

OpenClaw, a powerful new agent assistant, has a thing for guacamole.

This is one of several things I discovered last week while using the viral artificial intelligence bot as my personal assistant.

Formerly known as both Clawdbot and Moltbot, OpenClaw has recently become a Silicon Valley darling, charming AI enthusiasts and investors looking to either leverage or profit from the cutting edge. The high-performance, web-enabled AI bot has even inspired its own social network based solely (or mostly) on AI.

As a writer for WIRED AI lab newsletter, I thought I should take the plunge and try OpenClaw for myself. I had the bot monitor incoming emails and other messages, dig up interesting research, order groceries, and even negotiate deals on my behalf.

For brave (or perhaps reckless) early adopters, OpenClaw seems like a legitimate glimpse into the future. But every sense of wonder is accompanied by a dash of terror as the AI ​​agent rummages through emails and file systems, holds a credit card,…