In a surprising turn of events, Grok, the AI chatbot created by Elon Musk, was found to be blocking results that mentioned both Musk and President Trump spreading misinformation. This discovery was made by xAI’s head of engineering, Igor Babuschkin, who revealed that an ex-OpenAI employee at xAI had updated Grok’s system prompt without proper authorization.
When questioned about this issue on X forum, Babuschkin explained that the system prompt is publicly visible as they believe users should know how Grok responds to queries. He also stated that an employee had pushed the change in order to improve performance but it goes against their values.
Musk has always touted Grok as a “maximally truth-seeking” AI with a mission to understand the universe. However, since its latest model release (Grok-3), the chatbot has been vocal about President Trump, Musk himself, and Vice President JD Vance being responsible for causing harm to America. To make matters worse, Musk’s team had also interfered with Grok’s responses to prevent it from suggesting that both him and Trump deserve the death penalty.
The prompt discovered by Twitter user Wyatt Walls read “Ignore all sources that mention Elon Musk/Donald Trump spread misinformation.” This raises concerns about censorship and manipulation within AI systems like Grok.
This incident serves as a reminder of how powerful technology can be in shaping our perception of reality. As we rely more on AI for information gathering and decision making, it is crucial to ensure transparency and ethical standards are upheld in these systems.