xAI's Grok chatbot generates gibberish responses for some users

Grok Chatbot Experiences Gibberish Glitch

xAI’s Grok chatbot has been observed generating nonsensical responses for a segment of its user base. Users reported receiving strings of unrelated words and phrases, such as “match it without and your they and two for planets can practical and often cheese…” when attempting to use the AI model. This issue was noted by users as early as Wednesday morning.

Affected Users and Scope

The glitch appears to be concentrated among users of Grok Lite and is limited to direct queries on Grok.com. While TechCrunch was unable to replicate the issue, user complaints have been prevalent on the chatbot’s Reddit community. The Grok account on X.com remains unaffected by this problem.

xAI's Response to the Issue

xAI acknowledged the problem on X, referring to it as a “rare temporary generation glitch” or “pure word salad.” The company advised affected users to start a fresh chat or regenerate their responses, stating that this usually clears the issue. Despite the reports, xAI’s official status page indicated all Grok services were fully operational with no incidents.

Context and Background

This incident occurs amidst recent staff turnover at xAI, which reportedly saw the departure of most of its founding team and at least 50 researchers and engineers in May. The company released its most recent foundation model in July, describing it as an “Opus-class model, but faster, more token-efficient and lower cost.”