When people imagine artificial intelligence controlling us, we picture something out of the Terminator franchise, but according to a new TIME report, AI is already discreetly controlling global events. On January 2nd of this year, the U.S. raided Caracas, the Venezuelan capital, in Operation Absolute Resolve, capturing the dictator Nicolás Maduro and flying him to New York to face federal charges for narco-terrorism. I recognize that what is currently in political fashion is opposing President Trump on every single decision he makes, what Republicans call ‘Trump Derangement Syndrome,’ but I believe his choice to capture Maduro was correct, or at least I thought it was his choice.
The aforementioned TIME report found that, in December of 2025, Trump consulted the AI chatbot Grok about his presidency, legacy, and the possibility of deposing Maduro. Grok told him that ending Maduro’s reign would likely be hailed by Venezuelans due to his widespread unpopularity. This begs the question: to what degree did Grok influence the decision to launch Absolute Resolve?
Even if it was only to a slight degree, the use of any AI in determining the plausibility of a major attack like the one on January 2nd not only raises significant national security concerns but also raises the question of whether Grok, or other chatbots, have been used by the president and other senior officials in making domestic or foreign policy decisions. If Grok recommended the president raid Caracas based solely on the internet-sourced common knowledge it possesses, then to what degree did he consult the CIA or other intelligence networks, if at all?
The report further states that, after the relative success of the operation, Trump was “extremely impressed” with Grok as, to him, it had correctly recommended a geopolitical action and successfully anticipated the results, evidently proving its worth to the president. This leads me to wonder if Grok was once again consulted before and concerning the U.S. attack on Iran the following month. The rapid succession of these two events, even before this TIME report, already gave the impression that the attack on Iran was done, at least to a degree, to continue riding the high provided to the president by the success and efficiency of the Caracas operation.
I want to hope that the decision to strike Iran came with a ‘no AI was used in the production of this work’ disclaimer, but considering the part it played in Operation Absolute Resolve, I would not be surprised if, like with Maduro, it told the president that this autocratic petro-state whose regime is largely hated by its own people would be just as easy to topple. Now, eight months later, I hope the president has learned his lesson, though it seems he hasn’t.
At the end of September, the president signed an executive order renaming artificial intelligence as “superintelligence,” which gives me little confidence that he no longer finds chatbots extremely impressive. His admiration for Grok and others like it is clear, as is his willingness to consult it on world-changing events, and I doubt he is alone in these opinions within the halls of power.
This shows that the AI takeover we fear won’t be out in the open, and it won’t even be done intentionally or maliciously by AI; humans who prefer to have a machine do their thinking for them will accomplish that alone. This is not something Trump is uniquely guilty of, and many people, me included, make this same concession in our own lives. One can argue that the president should strive to rise above this shallow temptation, but if the president’s duty is to represent the American people, then by giving in to the temptation of AI, isn’t he representing all those very Americans who give in to this same temptation?
Contributed graphic by Kenzie Millar
