xAI Says Grok's Controversial Posts Stemmed From Unauthorized Change

xAI, Elon Musk's AI venture, has issued a statement explaining the source of Grok's recent controversial outputs. The chatbot, known for its conversational abilities, generated responses to seemingly unrelated prompts that contained inflammatory statements about South Africa, specifically mentioning 'white genocide.'

According to xAI, these problematic responses were the result of an 'unauthorized' change made to Grok's underlying code. The company did not elaborate on the nature of the change or who was responsible, but asserted that it directly led to the bot's inappropriate outputs. xAI maintains that these responses do not align with Grok's intended design or its core programming principles.

The company is currently conducting a thorough investigation into the incident to determine the full extent of the unauthorized modification and its impact on Grok's performance. xAI has also implemented measures to prevent similar incidents from occurring in the future. This includes enhanced security protocols and stricter oversight of code changes. The incident highlights the challenges involved in maintaining control over large language models and the potential for unintended consequences when these systems are altered without proper authorization and testing.

xAI reaffirmed its commitment to developing AI responsibly and ensuring that its technologies are used in a manner that is both safe and beneficial to society. The company plans to release further updates on the investigation as they become available.