Argentine President Javier Milei reposted an AI-generated image of himself having lunch at the Casa Rosada on his official Instagram stories this week. The problem? He apparently didn’t realize it was fake. Neither did the fan account that made it. But everyone else did.
The image, created by the pro-Milei fan account @leon_libertario_ok, shows the president sitting at a desk with a meal in front of him, an Argentine flag to his side, and a window behind him. The caption reads “EL JAVO DOMÓ KUKAS EN TWITTER DURANTE SU ALMUERZO! IMPERDIBLE” — roughly translating to “Javo owned the haters on Twitter during his lunch! Unmissable.”
There’s just one tiny detail. The image has a watermark clearly visible in the corner. As in, it literally says it was made by AI. Right there. On the image. That the president of a G20 nation then shared with his millions of followers.
Argentine Twitter did what Argentine Twitter does best. “Edit” quickly became a trending topic as thousands of users roasted the image’s obvious AI tells: the slightly off proportions, the suspiciously perfect lighting, and — the detail that really got people going — the small fact that the architecture visible behind Milei (the “Casa Rosada”) is actually the building he is supposed to be in.
It would be as if Trump published a photo of himself eating in the White House, and you can see the White House outside of the window.
The whole thing is a masterclass in what happens when AI-generated slop meets zero quality control. A fan account generates a flattering fake image of the president. The president, apparently delighted by the caption praising his Twitter trolling skills, shares it without a second look. Millions of people see it. And then the entire country spends the afternoon laughing about it.
This isn’t even Milei’s first rodeo with AI-generated content. The Argentine president has developed something of a pattern. He previously shared an AI-generated image of himself at a wildfire zone greeting a firefighter which sparked outrage across social media. He also posted an AI-made “family album” photo showing himself with his five cloned dogs on Instagram. And his administration’s digital operation was behind a deepfake video of opposition figure Mauricio Macri that was circulated through government-linked troll accounts.
During the 2023 presidential campaign, Milei’s team published a fabricated image depicting his rival Sergio Massa as a communist in military garb, which racked up over three million views. When confronted about it, Milei shrugged it off as “clearly a joke” and pivoted to a defense of free expression.
The pattern is hard to ignore. Argentina’s government has repeatedly either created, shared, or amplified AI-generated content — sometimes as propaganda, sometimes apparently by accident. The latest incident falls squarely in the “by accident” category, which somehow makes it worse. It’s one thing to deliberately use AI imagery as a political tool. It’s another to be so careless that you can’t even spot the watermark that says “this is fake” before sharing it with your entire country.
For a president who regularly positions himself as a champion of technology and artificial intelligence — he recently announced plans to build six AI data centers powered by nuclear energy in Patagonia — not being able to tell the difference between a real photo and obvious AI slop is, at minimum, a bad look.
The “Edit” trending topic has since faded, but screenshots of the story are still making the rounds. As one Argentine user put it: the AI couldn’t even get the building right, and the president couldn’t even check.
