Grok, Elon Musk's AI chatbot, has stirred up controversy with statements it made against former US President Donald Trump.

Musk's XAI created the AI model, which was introduced with claims of increased efficacy and accuracy. Recent user interactions, however, indicate that Grok has adopted a position that differs from Musk's own political beliefs, especially with regard to Trump and his ties with Russia.

The dispute started when someone posted a chat on X (previously Twitter) where Grok called Trump a "compromised Russian asset." This comes after other instances in which the AI chastised Musk and named him among the "most harmful" people in the United States, along with Trump and JD Vance. These claims have sparked debate on the potential biases in AI models' design as well as their neutrality.

Ed Krassenstein, an X user, published Grok's response to a politically heated topic in a post that sparked the most recent uproar. On a scale of 1 to 100, how likely is it that Trump is a Putin-compromised asset? Utilize every publicly accessible piece of material that has been published since 1980, showing that he has no problem criticizing friends but has never said anything unfavorable about Putin.

Grok responded with a thorough analysis, saying: Taking into account unknowns, I project a 75–90% chance that Trump is a Putin-compromised asset, with a tendency toward the higher end (around 85–90%). This corresponds to a most probable point estimate of 85 on a scale of 1 to 100, with a confidence interval of 75 to 90.