The Grok chatbot sent meaningless responses to some users
xAI’s Grok chatbot sent nonsensical responses to some users in response to direct queries on Grok.com. TechCrunch reported the issue, citing users of the service who encountered the glitch on Wednesday morning.
Specifically, after asking to create a PDF, one user received several paragraphs of random words. Another user, while checking the links in the response, saw a series of links to websites featuring research on reinforcement learning. The users who reported the issue to TechCrunch were using the Grok Lite version.
The TechCrunch editorial team was unable to reproduce the glitch during its own testing and noted that it likely affected only a small fraction of users. A significant number of complaints appeared in the Grok community on Reddit. Refreshing the session often restored the chatbot to normal operation, although some users reported that the problem persisted even after several refreshes.
For more breaking news, follow the UA.News Telegram channel.
Grok’s X account confirmed the issue, describing it as a rare, temporary glitch in response generation. They advised users to start a new chat or regenerate the response. On the xAI status page at the time, all Grok services were listed as operating without incidents. According to TechCrunch, the outage was likely limited to direct queries on Grok.com, while the Grok account on X was unaffected.
xAI did not respond to the publication’s request for comment. In July, the company unveiled its latest foundation model, calling it an Opus-level model but faster, more token-efficient, and cheaper.
