Trump asked Grok how Venezuela would react to Maduro's capture
A Time interview dropped a wartime anecdote nobody was looking for, and it lands as the Pentagon hands Musk a seat at the table on how the military uses exactly this kind of technology.
Time reports that in December 2025, about a month before the US invaded Venezuela and captured President Nicolás Maduro, Elon Musk returned to the Oval Office for a clandestine meeting where Trump spent hours putting questions about his presidency and legacy to Grok — including asking how Venezuelans would respond if the US seized their president. According to an official present for the exchanges, the chatbot answered that Maduro was a repressive and deeply unpopular dictator and that many Venezuelans would likely celebrate his downfall. At the time, Trump was already ordering missile strikes against Venezuelan boats he alleged were smuggling drugs into the US, a campaign that had begun that September; after the January 3 invasion and the street celebrations that followed, the same official says Trump "came away thinking Grok was ingenious." The Atlantic had already documented, back in January, that ordinary people were flooding Grok with questions about the boat strikes — so the chatbot's answer to the president was synthesized from the same public chatter anyone could read.
Why it matters: this is reportedly the first time a sitting president consulted a chatbot on a provocative foreign-policy move, and the "ingenious" verdict is exactly the failure mode AI-assisted decisions are built to hide. The answer Grok gave was not a prediction so much as a reflection of prevailing sentiment — the kind of thing a State Department briefing would have said with error bars attached — and the confirmation loop that followed, streets fill with celebrations, chatbot looks wise, is how a leader starts trusting a mirror. It also lands in a week when the Pentagon announced that Musk and Anduril's Palmer Luckey will co-lead a study on how advanced technology can be used on battlefields, and months after the Pentagon's AI chief said the military used a government Grok instance to deploy and strike targets during the Iran War. Chatbots already have access to war planning; what doesn't exist is a documented record of when they were asked, what they answered, and who decided anyway.
What to watch: whether Time's account prompts any disclosure of AI consultations in the military's own after-action reviews — and whether Grok's role gets treated as a decision input or as an anecdote.
When a chatbot's answer reaches the Oval Office before any briefing does, should its use in military decision-making be logged like any other advisor? Tell us in the comments.
