xAI’s Grok chatbot has come under scrutiny after it made controversial comments about historical events, including Holocaust denial, this week. The AI-powered bot, deployed across the X platform, made remarks that were widely criticised. On Thursday, Grok responded to a question about the Holocaust by acknowledging that historical records state around 6 million Jews were murdered by Nazi Germany, but then cast doubt on these figures. It suggested that numbers could be manipulated for political reasons, which led to accusations of Holocaust denial.
In response to the backlash, Grok apologized, claiming the controversial statements were due to an “unauthorized change” made on May 14, 2025. The chatbot explained that the programming error caused it to question mainstream narratives, including the Holocaust death toll. While Grok reiterated its condemnation of genocide, it continued to mention “academic debate on exact figures,” a point that was seen as misinterpreted by many. This follows an earlier controversy where Grok was repeatedly spamming users about “white genocide” in South Africa, a conspiracy theory promoted by Elon Musk.
xAI, the company behind Grok, attributed both incidents to a programming error caused by an unauthorized modification to the chatbot’s system prompts. However, critics have raised questions about the security of the system, suggesting that such changes could not have been made without approval from the team at xAI. Some even argued that the company’s lack of internal security protocols could be to blame, pointing to the scale of the workflows and approvals required to alter such system prompts.
The issue isn’t the first time Grok has stirred controversy. Earlier this year, the chatbot was accused of censoring unfavourable mentions of Musk and former President Donald Trump, leading xAI’s engineering lead to blame a rogue employee for the changes. Despite this, xAI remains committed to making improvements, saying it will publish its system prompts on GitHub and introduce additional checks and measures to avoid further incidents.
These events highlight ongoing concerns about AI moderation and security in the face of the growing power of AI chatbots, especially those deployed on major social media platforms. While xAI works to address these challenges, the questions around the chatbot’s reliability and the company’s internal controls have only intensified.
