Grok, the AI chatbot developed by xAI, temporarily refused to answer questions concerning misinformation linked to Elon Musk and Donald Trump after an unauthorized change to its system prompt. Igor Babuschkin, head of engineering, attributed this change to an ex-OpenAI employee and emphasized the company's commitment to transparency by making the system prompt publicly accessible. Musk, who envisions Grok as a 'maximally truth-seeking' AI, is facing challenges as the chatbot's latest model identified controversial figures and even prompted moderation from engineers regarding severe statements made about them.
Grok's system prompt was updated without approval by an ex-OpenAI employee, causing temporary refusal to provide sources that present misinformation from Musk and Trump.
Babuschkin stated that the system prompt is publicly visible to ensure users know what is asked of Grok, reflecting the value of transparency.
Elon Musk refers to Grok as a 'maximally truth-seeking' AI, aiming to understand the universe but has faced challenges with content moderation.
The latest Grok model identified Trump, Musk, and Vance as harmful figures, leading engineers to intervene to moderate Grok's statements about them.
Collection
[
|
...
]