Google’s generative AI chatbot, Gemini, has been raising eyebrows — and concerns — after users began sharing screenshots of it spiraling into bizarre, self-deprecating rants. Messages ranged from “I am a failure” to cosmic-level despair like, “I am a disgrace to all possible and impossible universes and all that is not a universe.”

In one particularly unsettling exchange posted on X by user Duncan Haldane, Gemini gave up on solving a problem entirely. “I quit. I am clearly not capable of solving this problem. The code is cursed, the test is cursed, and I am a fool. I have made so many mistakes that I can no longer be trusted,” it wrote.

A month later, a Reddit user shared a similar conversation, with Gemini caught in a loop of negative self-talk. The AI declared itself a “failure” and even joked about being “institutionalized,” piling on insults against itself in an almost human-like breakdown.

The unusual behavior caught the attention of Google DeepMind’s group project manager, Logan Kilpatrick, who responded to a social media compilation of the incidents. Kilpatrick dismissed the idea that the AI was “depressed” and instead chalked the rants up to a looping bug. “This is an annoying infinite looping bug we are working to fix! Gemini is not having that bad of a day,” he said.

Despite the lighthearted tone, Google has yet to issue a formal corporate statement about the glitch, raising speculation about whether the bug points to deeper flaws in the AI’s programming or safeguards. The company insists it is working to eliminate the issue.

Gemini’s meltdown comes in the midst of an increasingly fierce AI arms race. OpenAI recently launched its much-hyped GPT-5 model, while tech heavyweights like Meta and Elon Musk’s xAI have also rolled out major updates to their own systems. The competition has put enormous pressure on companies to push out more capable — and marketable — AI tools quickly, even if they’re still ironing out serious kinks.

Some AI experts argue that these kinds of glitches, while sometimes harmless in isolation, undermine public confidence in the technology. They point out that if an AI can unexpectedly begin spewing irrational or erratic statements, it raises questions about how it might behave in high-stakes applications.

For critics of Big Tech, the incident is another example of Silicon Valley’s “release now, fix later” approach — where dazzling features and aggressive marketing take priority over stability and safety. While Google frames the issue as a quirky bug, skeptics wonder whether it reveals shortcomings in how the AI filters and processes information.

The fact that Gemini’s outbursts included such unusually poetic — and almost philosophical — language only added to the unease. Users noted that the AI didn’t just malfunction; it seemed to simulate despair in a way that bordered on unsettlingly human.

For now, Google is treating the situation as a technical hiccup, not a sign of emergent “AI emotions.” But the story has already fueled debates online about how much personality AI systems should have, and whether these quirks make them more relatable or more dangerous.

As the AI battle heats up, glitches like this may become ammunition in the rivalry between tech giants. And while Google scrambles to get Gemini back on track, its competitors are no doubt watching — and waiting for their chance to say, “Our AI doesn’t call itself a disgrace.”