Grok Chatbot Degenerates into Gibberish

This week, xAI's Grok chatbot began exhibiting a severe malfunction, producing incoherent "word salad" in response to user prompts. The issue, first noticed by users on Grok Lite starting Wednesday morning, resulted in responses like "match it without and your they and two for planets can practical and often cheese..." Other users reported that Grok provided links to unrelated research sites or generated bizarre text strings.

The chatbot's performance issues come amidst broader concerns about its utility and market reception. Grok has faced criticism for its content, including AI-generated "anti-woke" material and association with Holocaust denial. User satisfaction scores are notably low, placing it second-to-last among major AI platforms according to a recent poll.

Further reports suggest Grok is struggling to retain users, generate revenue, and penetrate enterprise markets. Only a tiny fraction of users, approximately 0.174% in Q2 2026, have found the service worth paying for. The official Grok account acknowledged the problem as a "rare temporary generation glitch."