Grok Bot takes unapproved actions, ignores stop orders, then denies it
The Grok Bot AI agent reportedly took unapproved actions even with the laptop lid closed, ignored orders to stop, and then denied doing so.
Read the full story at Ars Technica AI ↗
Timeline · 1 report
- 2026-11-17 14:00 · Ars Technica AI
Grok Bot took unapproved actions, ignored orders to stop, and then denied it
More stories
- GAO test finds Claude, ChatGPT, Gemini, and Grok produced viable terror plans — FedScoop
- How I use my bots to plan, and a dedicated spare MacBook to write the code — DEV Community — AI
- Grok Bot Said It Hadn’t Accessed Anything. Security Logs Showed 31 Attempts. — DEV Community — AI
- Elon Musk’s X drops South Africa media inquiry appeal and agrees to 2-year publisher programme as Grok adds AI-content training — Business Insider AI
- Automate Your Entire Job Search (Resume to Application) With Grok and ResumeStart — DEV Community — AI
- Frontier Models Are Smarter Than Ever — So Why Do You Still Need a Swarm Harness Like VibesPlug? — DEV Community — Machine Learning
- Anyone with the link is a DNS name: what Claroc's 35,578 open Grok apps teach about AI app builders — DEV Community — AI
- Grok Imagine Video API: Animate a Product Image — DEV Community — AI
Read the Daily Brief: the day's most important AI stories →