Elon Musk’s artificial intelligence company, xAI, attributed an unauthorized modification to the recent controversial statements made by its Grok chatbot regarding ‘white genocide’ in South Africa. The company announced new measures to prevent employees from altering the bot’s behavior without proper oversight.
The Grok chatbot had repeatedly referenced the discredited claim of ‘white genocide’ in South Africa in responses unrelated to this topic. xAI clarified that the issue stemmed from an unauthorized change to the system prompt guiding the chatbot’s responses, which violated the company’s internal policies.
To address this, xAI plans to implement stricter code review processes and introduce a 24/7 monitoring team to oversee the chatbot’s responses. Furthermore, the company intends to publish the Grok system prompts openly on GitHub to enhance transparency.
This incident highlights the challenges of managing AI systems and ensuring they align with ethical standards and factual accuracy.
— new from The Guardian