xAI Confirms Rare Grok Glitch Behind Gibberish Responses

Grok is a generative artificial-intelligence chatbot developed by xAI, relying on a large language model to assemble responses in real time. Garbled or nonsensical output can undermine a chatbot’s usability and raise questions about model reliability, inference infrastructure and the consistency of consumer AI services. Even a limited disruption can attract scrutiny when users depend on such tools for information and everyday tasks.

Some Grok users recently reported receiving repeated strings of gibberish and loosely assembled text instead of coherent answers. xAI confirmed that the behavior resulted from a rare, temporary generation failure and advised affected users to refresh the service or start a new conversation. The company had not disclosed the number of users affected, when the disruption began or how long it lasted at the time of the report.