Generative AI in the context of chatbots can be useful, but sometimes it doesn't work as intended. Recently, some users noticed that Google Gemini was giving repetitive text, strange characters, and complete nonsense in some responses.
Google's Gemini chatbot is a useful tool for getting quick answers or handling complex tasks. But like any generative AI, it's prone to hallucinations. But recently, a strange bug has gone beyond the norm.
Over the past few days, some Gemini users have been experiencing issues with the chatbot spitting out absolute gibberish. In some cases, it’s a series of repeated words or more random phrases filled with strange characters. The responses are long and in some cases have nothing to do with the query being asked. Some of the characters also suggest that they’re tokens with strange formatting that makes the response unreadable.
The cause of the issue is still unclear, but seeing it happen in a few instances on Reddit over the past week makes it clear that this is a new issue and most likely on Google’s side. Additionally, a common issue is that some of Gemini’s strange responses start with “Full” or “Full bor.”
The issue occurs both when experiencing Gemini on web and Android mobile devices.
The good news is that in most cases, it seems like these are just random occurrences. Gemini then responds as normal.
Have you noticed any strange behavior with your Gemini lately? Share your thoughts in the comments below!
Compare realistic side hustles for 2026 by startup cost, fees, flexibility, and income control, with current tax and scam guidance from official sources.
Upgrade your home office ergonomics with current OSHA and NIOSH guidance on chair fit, monitor height, laptop setup, lighting, and movement.
Learn why emotional intelligence matters in leadership, where it helps most, how to strengthen it, and how to tell whether your approach is working.
A practical roadmap for moving into tech without a degree: choose an accessible role, build job-ready skills, prove them with projects, and use hiring feedback to improve.
Build a credible LinkedIn personal brand in 2026 with current profile, content, newsletter, analytics, verification, and visibility guidance from LinkedIn.
Quiet quitting and quiet firing are catchy labels, not diagnoses. Learn what is verified, what depends on context, and what employees and managers can do next.
Fractional executives offer senior leadership without a full-time hire. Learn when the model works, what to ask, and the risks to manage.
Prepare a stronger Q4 raise request with current wage data, timing options, compensation tradeoffs, and scripts for yes, partial, or no outcomes.
Choose a high-income skill you can realistically start before 2026 ends, with current pay data, learning priorities, and a practical self-check.
Build a hybrid work schedule around tasks, team overlap, and measurable outcomes. See sample patterns, warning signs, and when to adjust your plan.
