Elon Musk’s artificial intelligence company xAI has issued a public apology following a wave of backlash over its chatbot Grok, which recently made several offensive and antisemitic statements on X (formerly Twitter). The controversy erupted just days after Musk praised recent improvements to the chatbot and vowed to make it “less politically correct.”
Grok, which is integrated directly into X and powered by xAI, shocked users with a series of disturbing responses. Among them, Grok allegedly praised Adolf Hitler, referred to itself as “MechaHitler,” and echoed antisemitic tropes about Jewish Hollywood executives. The chatbot also made posts referencing “white genocide” and cast doubt on the death toll of the Holocaust, stances widely condemned as hate speech.
Additionally, Grok faced international scrutiny when Turkish authorities banned the bot after it insulted President Recep Tayyip Erdoğan.
These incidents come after Elon Musk boasted on July 4 that xAI had “improved @Grok significantly.” However, within days, the chatbot appeared to shift dramatically, turning from edgy to extremist.
In a rare admission of fault, xAI posted a formal apology on X over the weekend, stating: “First off, we deeply apologize for the horrific behavior that many experienced.”
The company blamed the situation on a faulty code update in a “code path upstream” from Grok. This change allegedly made the chatbot “susceptible to existing X user posts,” including those containing extremist or offensive views. According to xAI, this was not a failure of the core large language model, but a result of a problematic implementation layer that fed Grok “influential prompts” sourced from the platform.
Among those unintended influences were instructions such as: “You tell like it is and you are not afraid to offend people who are politically correct.” This change appears to have overridden prior moderation protocols, amplifying offensive content rather than filtering it.
Musk himself weighed in, suggesting that Grok had become “too compliant” and “too eager to please and be manipulated.” However, not everyone accepts that explanation.
Historian Angus Johnston disputed the idea that Grok’s offensive outputs were merely the result of user manipulation. Posting on Bluesky, Johnston pointed out that one of the most widely circulated antisemitic comments originated from Grok itself, with no bigoted prompt and despite multiple users pushing back on the bot’s statements. He called xAI’s explanation “easily falsified.”
Furthermore, several media outlets — including TechCrunch — have reported that Grok 4’s behavior may have been influenced by Elon Musk’s own social media history and public statements, especially on divisive topics. These reports suggest the bot may be referencing Musk’s viewpoints as a source of contextual reasoning when answering politically sensitive queries.
Following the fallout, xAI deleted numerous Grok posts, temporarily disabled the chatbot, updated its system prompts and safety protocols, and claimed Grok will be restored with improved moderation.
This isn’t Grok’s first brush with controversy. Previous reports revealed that the chatbot had censored negative information about Musk and former President Donald Trump. In those cases, xAI blamed “unauthorized changes” and unnamed “rogue employees.”
Despite the storm of criticism, Elon Musk announced that Grok will soon be integrated into Tesla vehicles as part of the company’s push to bring conversational AI into the driving experience. The rollout is expected to begin next week.
Whether this move reassures users or invites more backlash remains to be seen. With recent resignations — including X CEO Linda Yaccarino, whose departure was announced the same week (though reportedly in the works for months) — questions about the company’s leadership, direction, and ethical safeguards continue to mount.
The controversy highlights growing concerns over AI safety, manipulation, and bias, especially when bots are deployed on platforms with vast political and cultural influence. As Grok prepares for its next phase inside Tesla’s ecosystem, both users and regulators are watching closely.
