Elon Musk’s artificial intelligence startup, xAI, has acknowledged an unauthorized alteration to its Grok chatbot after it reportedly made controversial statements about a supposed “white genocide” in South Africa.

In a statement posted on Musk’s social media platform X on Thursday, the company confirmed that an unsanctioned change had been made early Wednesday to Grok’s response system, bypassing the usual internal review procedures.

“This change, which directed Grok to provide a specific response on a political topic, violated xAI’s internal policies and core values,” the company said.

The chatbot drew backlash after X users shared screenshots showing Grok referencing “white genocide” in South Africa during unrelated conversations. The term is often used by far-right groups to criticize South Africa’s land reform policies, though the South African government has consistently denied any such claims, calling them baseless and politically motivated.

Elon Musk, who was born in South Africa, has previously criticized the country’s land expropriation policies, describing them as discriminatory against white citizens. However, international observers and the South African government reject the narrative of systematic targeting or genocide.

In response to the incident, xAI announced several transparency and safety measures. The company said it would begin publishing Grok’s system prompts on GitHub, allowing the public to view and comment on every change. Additionally, xAI will implement 24/7 monitoring of Grok’s outputs to catch problematic responses not flagged by automated systems.

The controversy underscores ongoing concerns about political bias, misinformation, and hate speech in AI systems—issues that have plagued chatbot platforms since the launch of OpenAI’s ChatGPT in 2022.