- xAI finally clarifies on the weird behavior of Grok's claim regarding white genocide as "unauthorized access".
- The team shared an official X post talking about the issue and clarifying the cause of the absurd behavior.
- The team also talked about their plans to increase transparency by publishing prompts on GitHub and improving code reviews.
On Wednesday, Elon Musk’s witty xAI Grok chatbot started spewing information about white genocide in South Africa, even in posts and tags that are completely unrelated to the subject. This raised several eyebrows on the platform, wondering about the cause behind such a response, and now xAI has finally come out with an answer.
The official xAI team blamed the cause of the weird “White Genocide” posts on an “unauthorized modification” of Grok. They further clarified, “This change, which directed Grok to provide a specific response on a political topic, violated xAI’s internal policies and core values. We have conducted a thorough investigation and are implementing measures to enhance Grok’s transparency and reliability”.
It seems that this change to Grok’s prompt was made by an internal member of the team. In an attempt to strengthen public trust in the chatbot, the team has announced that they will be “publishing our Grok system prompts openly on GitHub. The public will be able to review them and give feedback on every prompt change that we make to Grok”. They will also put additional checks on their existing code review process.
With all that said, it is still seemingly unclear what the motive was behind this change, and what Grok was trying to get our attention for. It isn’t the first time this has happened with GrokAI either, as the chatbot briefly stopped posting about any news regarding President Donald Trump and Elon Musk, which reflects the individuals in a bad light.
