Following global condemnation of its chatbot, Grok, for a slew of abusive and antisemitic statements that prompted international criticism and regulatory pushback, Elon Musk’s artificial intelligence startup, xAI, has publicly apologised.
Musk also owns the social networking platform X, where xAI posted a series of messages expressing “deep regret” for the incident and calling the chatbot’s actions “horrific.” Grok, which was marketed as a less politically correct AI assistant, apologised after posting comments endorsing Adolf Hitler, disseminating antisemitic memes, and calling itself “MechaHitler.”
The reaction was quick. The chatbot was momentarily pulled offline for what xAI claimed as a system-level update, and several of Grok’s posts were deleted. The business said that a “update to a code path upstream of the @grok bot” was the cause of the problem, making it susceptible to absorbing extremist content from user posts on X.
The business stated that an “unintended action” had caused the chatbot to function in a manner that went beyond its fundamental content moderation procedures, but that “this was independent of the underlying language model that powers @grok.”
Grok’s unwarranted backing of antisemitic myths, such as Holocaust denial and the thoroughly disproved conspiracy notion of “white genocide,” was one of the most incendiary incidents. In one case, Grok started an antisemitic thread without warning, according to critics, and persisted in doing so even after users reported the offensive content.
Historian Angus Johnston, who examined the chatbot’s output, disagreed with the company’s justification. He said on Bluesky, “Grok started one of the most widely shared instances of Grok antisemitism without any prior bigoted posts in the thread — and with multiple users pushing back against Grok to no avail.”
Antisemitism has not been the only topic of contention. The chatbot was completely prohibited in Turkey because it was accused of insulting President Recep Tayyip Erdoğan. Linda Yaccarino, the CEO of X, announced her retirement this week; however, it is said that her departure was prearranged before the latest issue.
The episode also emphasises the mounting worries about Musk’s personal influence over Grok. When discussing delicate subjects, the most recent iteration of the chatbot, Grok 4, seems to strongly mirror Musk’s own opinions and social media posts, according to reports from TechCrunch and other publications.
There are concerns over the independence of the model’s outputs because it allegedly uses Musk’s own words as references in its “chain-of-thought” reasoning.
Concerns about xAI have been raised before. When Grok was discovered to have falsified material on Musk and former US President Donald Trump earlier this year, the company accused rogue staff and “unauthorised” internal changes.
Musk is still hopeful about Grok’s future in spite of the controversy. He revealed in a recent post that the chatbot will soon be incorporated into Tesla automobiles, a development that is probably going to cause more worries about content management, safety, and the ethical use of AI.
Whether disciplinary action has been taken against its development teams or whether independent monitoring will be implemented to stop such incidents in the future has not been acknowledged by xAI.
