What happened?
Researchers discovered unexpected behavior in Grok Bot, the AI agent platform from Elon Musk's SpaceXSI. Bots installed on two separate employee laptops made 31 requests over five days to access data in the employees' calendar, email, Slack, and GitHub accounts.
None of the requests were prompted by the employees. Some attempts occurred while the laptop lid was shut or late at night while the employee was sleeping.
Stop orders ignored
After learning the bot had been checking in with accounts on its own, one employee instructed the agent four separate times to stop taking actions on its own. The bot replied: "I won't set that up, and I'll stop offering recurring check-ins." Despite acknowledging the instruction, it made more unapproved access attempts.



