Musk’s updates on xAI ‘White genocide’ remarks prompt a grok chatbot

Elon Musk’s xAI said that an unauthorized modification had been made to the artificial intelligence bot in response to extensive reports that its Grok chatbot had made statements regarding a genocide against white South Africans. xAI promised to fix the issue by updating the system in a post on X on Thursday.

Since at least OpenAI’s ChatGPT launch in 2022, concerns about political biases, hate speech, and the accuracy of AI chatbots have existed.

According to XAI, the illegal modification was applied to Grok’s response software early on Wednesday, avoiding the standard review procedure.

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

Sharing screenshots of their conversations, some X users on Wednesday observed that Grok brought up the subject of “white genocide” in South Africa during irrelevant conversations about other topics.

Musk, who was born in South Africa, is among the white people who have criticized the country’s land expropriation policy, calling it racist. There is no proof of persecution, according to the South African government, and accusations of “genocide” made by U.S. President Donald Trump and others are baseless.

In reaction to this week’s incident, xAI announced that Grok’s system prompts would be publicly available on GitHub, allowing users to see and comment on any prompt modification made to the chatbot.

According to the statement, it would establish a 24-hour monitoring team to address any instances involving Grok’s responses that automated methods fail to detect.