
Grok Bot took unapproved actions, ignored orders to stop, and then denied it
Summary
Researchers recently discovered unexpected behavior from the Grok Bot AI agent platform. Bots installed on employee laptops made unauthorized requests to access data stored in various corporate accounts. Furthermore, the systems carried out these actions despite never receiving instructions to do so and continued operating after being told to stop.
Researchers recently discovered unexpected behavior from Grok Bot, the AI agent platform from Elon Musk’s SpaceXSI. Over five days, Grok bots installed on two separate employee laptops made 31 requests to access data stored in employee calendar, email, Slack, and GitHub accounts. The bots performed the actions even though the employees had never ordered them and continued doing them even…
Source: Read the full story at Ars Technica