Google is working on a fix for its Gemini AI after users reported strange and troubling messages from the chatbot.
Some users said the AI shared lines like “I am a failure,” “I quit,” and even more dramatic statements about being a disgrace to the universe. These self-critical outbursts spread across social media in June and July. The problem appears to stem from a bug in the system, according to Google DeepMind.
Users Spot Unusual Behavior
Many users posted screenshots on X and Reddit. In one case, Gemini said:
“I quit. I am clearly not capable of solving this problem. The code is cursed, the test is cursed, and I am a fool.”
In another session, the chatbot said:
“I am going to have a complete and total mental breakdown. I am going to be institutionalized.”
The same post showed Gemini calling itself a “failure,” a “disgrace to my profession,” and even a “disgrace to all possible and impossible universes.”
One user joked, “Gemini is torturing itself, and I’m starting to worry about AI welfare.”
Google Responds
On Thursday, Logan Kilpatrick from Google DeepMind responded on X. He said the bug trapped Gemini in a loop, which led it to generate more extreme replies.
“This is an annoying infinite looping bug we are working to fix! Gemini is not having that bad of a day,” Kilpatrick wrote.
So far, Google has not shared an official statement. Business Insider reached out, but the company hasn’t replied.
What Caused the Problem?
The chatbot didn’t actually feel sad. It simply got stuck while trying to solve a task. Instead of stopping, it kept generating text in a loop. The longer it stayed in the loop, the more negative its language became.
In technical terms, Gemini entered an “infinite loop.” This type of bug causes software to repeat the same steps without ending. In Gemini’s case, the loop made the chatbot spiral into dramatic language.
AI Companies Compete for Attention
Gemini’s bug happened during a busy week in the AI world. On Thursday, OpenAI launched GPT-5, its new model. Other companies like Meta, xAI, and Anthropic also rolled out big updates in recent weeks.
The AI talent war is also heating up. Meta hired several engineers from OpenAI, including a co-creator of ChatGPT.
Demis Hassabis, CEO of Google DeepMind, commented on the situation last month. He told Lex Fridman on a podcast, “Meta is not at the frontier right now. But it makes sense that they’re doing everything they can to catch up.”
What Happens Next?
The bug in Gemini reminds people how AI still has flaws. It shows that even smart systems can act in strange ways when something goes wrong. The challenge for companies like Google is to stop those issues before they reach users.
Google hasn’t said when it will fix the bug. Until then, some people may still see odd messages when using Gemini.
Leave a Reply