Elon Musk's xAI confirms an 'unauthorized modification' caused the chatbot's political outbursts. The issue highlights how AI systems are direct reflections of the people behind them.

Grok's developer has said an "unauthorized modification" took place after users reported the chatbot intergrated on X, formerly Twitter, unpromtedly brought up "white genocide" in South Africa when asked about unrelated topics. xAI said the change violated its "internal policies and core values."

Coincidentally, Musk, the leader of xAI and a Grok power user, has a known interest in the subject. In fact, he spent yesterday tweeting about white genocide in South Africa, his home country. In right-wing circles, the "white genocide" theory is a common talking point, although a South African court recently called the issue "clearly imagined."