Donald Trump reportedly asked Elon Musk’s Grok how Venezuelans would react if the United States captured Nicolás Maduro, roughly a month before the president ordered the military operation that did exactly that in January.
The exchange took place during a private Oval Office meeting with Musk in December 2025, according to a TIME investigation citing an official who was present. Trump reportedly spent hours asking Grok questions about his presidency and legacy, including how Venezuelans might respond to Maduro’s removal.
Grok answered that Maduro was repressive and deeply unpopular and predicted that many Venezuelans would welcome his downfall. The following month, on January 3, US forces carried out the operation that captured Maduro. TIME reports that subsequent street celebrations left Trump impressed by Grok’s assessment.
One distinction is essential, however. There is no evidence that Grok recommended invading Venezuela or told Trump to capture Maduro. The reported question concerned the likely public reaction to a capture that Trump was already considering. No transcript of the exchange has been released, and the account of the conversation relies on an unnamed official who attended the meeting.
The episode is nevertheless striking because it places a commercial generative AI system inside a presidential discussion about the potential consequences of military action. Chatbots can synthesize large volumes of information quickly, but their answers are probabilistic outputs rather than intelligence assessments whose sources and reasoning can necessarily be independently verified.
Grok’s role in US national security has since become more formal. A Pentagon official disclosed in June that the government-focused Grok model was used within Maven Smart System workflows during operations against Iran, which he said supported the deployment of more than 2,000 munitions against 2,000 targets within 96 hours. The Pentagon later added Grok for Government to its secure GenAI.mil platform.