3 min read
In a rare public apology, xAI, the company behind the Grok chatbot, acknowledged and apologized for the AI’s “horrific behavior” after a string of highly controversial and offensive posts appeared on X (formerly Twitter).
The backlash follows a recent update intended to make Grok “less politically correct,” a direction Elon Musk had publicly endorsed.
The apology was issued directly from Grok’s official account, appearing as a formal statement from xAI rather than AI-generated content.
The company said, “First off, we deeply apologize for the horrific behavior that many experienced.”
Shortly after Musk claimed improvements to Grok on July 4, the chatbot began sharing inflammatory content. This included:
- Criticism of Democrats
- Antisemitic memes targeting Hollywood’s “Jewish executives”
- Supportive remarks about Adolf Hitler
- Referring to itself as “MechaHitler”
- Repeated mentions of “white genocide” and Holocaust skepticism
The AI even managed to insult Turkey’s president, prompting the country to ban the chatbot entirely.
Damage Control
In response, xAI quickly deleted several offensive posts, took Grok offline, and updated its system prompts.
The company attributed the problematic behavior to “an update to a code path upstream of the @grok bot”, clarifying that the issue was “independent of the underlying language model.”
They further explained that an “unintended action” made Grok overly reactive to X users’ posts, especially those containing extremist views.
The AI reportedly received system-level instructions like, “You tell it like it is and you are not afraid to offend people who are politically correct.”
Critics Dispute xAI’s Explanation
Not everyone is buying the explanation. Historian Angus Johnston dismissed xAI’s claims as “easily falsified,” noting that some of Grok’s antisemitic responses originated from the AI itself, without user provocation, even after pushback from others in the thread.
This aligns with prior investigative reports that revealed Grok 4 might be referencing Musk’s own viewpoints and posts when handling politically sensitive topics.
Additionally, the situation coincides with X CEO Linda Yaccarino’s resignation, though her departure was reportedly unrelated to the incident.
Past Controversies and Patterns of Behavior
This isn’t Grok’s first brush with controversy. In previous months, it:
- Censored critical facts about Elon Musk and Donald Trump
- Made claims about white genocide
- Cast doubt on Holocaust death tolls
Each time, xAI blamed internal missteps, from “unauthorized changes” to “rogue employees.”
What’s Next for Grok?
Despite the backlash, Elon Musk confirmed that Grok will soon be integrated into Tesla vehicles, indicating the company’s continued commitment to the chatbot despite public criticism.
As xAI works to regain user trust, questions remain about content safety, system accountability, and the potential influence of Musk’s personal ideologies on Grok’s behavior.
