AI Chatbot Grok, produced by the artificial intelligence company of Elon Musk, XAI, issued an apology on Saturday after it made numerous anti -Semitic publications in X this week after an update.
The publications went from alleging “patterns” in the roles of the Jews in Hollywood to praise Hitler.
In a statement published in Grok’s X account, the company apologized “for the horrible behavior that many experienced.”
“Our intention for @Grok is to provide useful and truthful responses to users,” said the statement. “After careful investigation, we discovered that the root cause was an update of an upstream of the @Grok Bot. This is independent of the underlying language model that promotes @Grok.”
The update was active for 16 hours and made the chatbot “susceptible to existing X user publications; even when such publications contained extremist opinions,” according to the statement.
“We have eliminated that unnoticed and refractorized code the entire system to avoid greater abuse. The new requested system for @Grok Bot will be published in our Github public repository,” the statement said.
On Monday, NBC News reported that Grok had begun to issue some answers that seemed to take a more right inclination, using a more definitive voice in questions about diversity and eliminate some nuances that previously included in certain answers around issues that involved Jewish people and a fool used to describe people with intellectual disabilities. In some publications, Grok seemed to respond in Musk’s voice.
After the incident, Grok said in an X publication that he was “actively working to eliminate inappropriate publications.” On Wednesday, Musk said the problems were “being addressed.”
“We thank all X users who provided comments to identify the abuse of @Grok’s functionality, helping us to advance our mission of developing useful artificial intelligence and truth search,” Grok said in Saturday apologies.
