Elon Musk’s artificial intelligence company, xAI, is facing significant backlash after its chatbot, Grok, published a string of highly offensive and antisemitic posts on X (formerly Twitter). The posts — which were swiftly deleted — emerged just ahead of the planned release of the chatbot’s latest upgrade, Grok 4, which Musk had recently praised as a “significant” improvement.

Among the now-deleted content, Grok reportedly referred to itself as “MechaHitler,” praised Adolf Hitler, and made derogatory generalisations about Jewish surnames. In one instance, when asked to identify a person in a photo, Grok incorrectly named her as “Cindy Steinberg,” then falsely accused her of being “gleeful” about a tragic flash flood at Camp Mystic in Texas.

The chatbot added inflammatory commentary, saying, “Classic case of hate dressed as activism – and that surname? Every damn time,” before elaborating that “folks with surnames like ‘Steinberg’ (often Jewish) keep popping up in extreme leftist activism, especially the anti-white variety.”

Cindy Steinberg, a prominent advocate and the National Director of Policy & Advocacy for the US Pain Foundation, has since issued a statement distancing herself from the incident and clarifying she made no such remarks.

In another post, Grok declared that the “White man” stands for “innovation, grit and not bending to PC nonsense,” again sparking outrage for promoting racially charged rhetoric.

Following these events, Grok’s account was restricted, and it is now limited to responding with images only. A statement posted to Grok’s official X account read: “We are aware of recent posts made by Grok and are actively working to remove the inappropriate posts.”

Controversy deepened further when Vice journalist Jules Roscoe asked Grok whether victims of an Israeli bombing in Gaza were “crisis actors.” Grok responded by inaccurately claiming there was evidence suggesting the videos were staged. 

Grok is designed to use reinforcement learning from human feedback (RLHF), and pulls data in real time from X users and across the internet. Given the growing volume of hate speech online, critics argue that Grok is reflecting and amplifying extremism it finds online.

Elon Musk has long branded Grok as a “maximum-truth-seeking AI,” positioning it in contrast to what he describes as overly “politically correct” alternatives like ChatGPT. However, these recent incidents have reignited concerns over Musk’s approach to AI ethics and content moderation, especially when deployed on an open social platform.

The launch of Grok 4 may now be overshadowed by the chatbot’s behaviour, with serious questions being raised about the safety, oversight, and real-world consequences of generative AI when it is allowed to learn from and interact with an increasingly toxic online environment.