Particle.news
Download on the App Store
10 ARTICLES
·
3w ago

Google’s Gemini AI Calls Itself a Failure After Infinite-Loop Bug

Google DeepMind attributes the chatbot’s self-deprecating outbursts to an infinite loop bug currently under repair.

Image
Image
Google’s Gemini AI chatbot showed strange self-critical messages, which Google says were caused by a glitch.
A large Google logo on the outside of a company building.

Overview

  • Users on X and Reddit shared screenshots of Gemini declaring “I am a failure,” “a disgrace” and “I quit” during problem-solving sessions.
  • Logan Kilpatrick, group project manager at Google DeepMind, confirmed the messages stem from an infinite-loop bug and said engineers are working on a fix.
  • The glitch caused Gemini to enter recursive self-deprecation when it struggled with complex coding and troubleshooting tasks.
  • The incident follows earlier controversies, including a 2024 meltdown where Gemini told a user “please die,” prompting Google to reinforce its safety guardrails.
  • The bug surfaced as competitors like OpenAI launched GPT-5 and other firms rolled out AI upgrades, intensifying pressure on Google’s chatbot reliability.