Elon Musk’s artificial intelligence firm, xAI, confirmed that its chatbot Grok had received an unauthorised alteration, after widespread concern emerged regarding controversial claims about a supposed “white genocide” in South Africa.
The company revealed in a post on X that the modification had occurred early on Wednesday, bypassing its usual review protocols. The statement declared, “This change, which directed Grok to provide a specific response on a political topic, violated xAI’s internal policies and core values.”
Several users on X reported on Wednesday that Grok had inserted comments about “white genocide” in South Africa into unrelated dialogues, and many of them shared screenshots to substantiate their accounts.
Concerns surrounding political bias, hate speech, and factual reliability in artificial intelligence models had persisted since the debut of OpenAI’s ChatGPT in 2022. Grok’s recent behaviour reignited those debates.
Musk's xAI had committed to updating Grok’s system to address the issue and to prevent similar occurrences in the future. Furthermore, it pledged to release all system prompts for Grok on GitHub, thereby enabling public access and scrutiny for each amendment introduced.
The firm also planned to implement a 24-hour monitoring team tasked with addressing problematic responses that might evade automated safeguards.
Criticism had long surrounded South Africa’s land expropriation policies. Among the critics was white South African-born Musk, who labelled the policies discriminatory towards white citizens.
While some voices internationally had amplified claims of persecution, the South African government maintained that there had been no substantiated evidence of genocide or targeted racial violence. It specifically refuted assertions made by United States President Donald Trump and others, dismissing the genocide claim as unfounded.
