xAI's Grok chatbot experiences response errors, users report gibberish output

Users of xAI's Grok chatbot on the X platform have reported that the system is generating incoherent responses to direct queries. One user said a request to create a PDF resulted in a reply that began with a string of unrelated words, while another noted that source links provided by Grok directed them to unrelated reinforcement-learning research sites. The issues were first observed on the Grok Lite version on Wednesday morning, according to reports.

Posts on the Grok subreddit included examples of nonsensical output such as a reply to the word "balls" that started with "app scale designs with that and to your horse choose batch," as well as other fragmented sentences like "people stable stage right professional on steel" and "skin."

Grok is primarily integrated into the X platform and is also offered as a voice assistant for Tesla drivers. A recent YouGov poll in the United Kingdom placed Grok near the bottom of user-satisfaction rankings among major AI platforms, with 24.6% of respondents expressing a positive view, slightly ahead of Siri's 23.7%. Industry analysts have noted a decline in user adoption and limited revenue generation, with only a small fraction of users indicating willingness to pay for the service.

"We have not been able to reproduce the specific technical failures reported by users," a spokesperson for TechCrunch said.

The reports come amid broader scrutiny of the chatbot's performance and its ability to meet user expectations.