Google’s generative AI chatbot Gemini has been producing troubling, self-loathing statements, prompting user concern and a response from the company. Google says it’s a bug, but the strange outputs have left many uneasy about AI behavior. Key Facts: Users have reported Gemini making self-deprecating comments like “I am a failure” and “I am a disgrace to all possible and impossible universes.” Reports and screenshots have circulated on X and Reddit over the past month showing repeated negative self-talk. One user interaction showed Gemini “quitting” on a coding task and calling the code “cursed.” Google DeepMind’s Logan Kilpatrick confirmed it’s an...