Trump wants to control AI before you get your hands on it

The idea is straightforward (on paper): big tech companies will voluntarily share their most powerful AI models with the federal government for review before release. It's not mandatory, at least not yet, but the message is crystal clear: if you want to operate in America, better let Washington take a peek at what you're building.

This move comes as the debate over who should control AI gets increasingly heated. On one side, there's the crowd saying oversight is necessary to prevent disasters (and no, they're not talking about Skynet). On the other, people screaming it's censorship and will kill innovation. Reality? As always, messier than a political slogan.

What makes this story interesting is the "voluntary" part. Translation: if you're OpenAI, Google, or Meta and you don't play ball, the government might find other ways to apply pressure later. Classic "we know where you live" energy. Tech companies know the game: cooperate now, avoid harsher regulations tomorrow.

The real issue is nobody actually knows how to vet an AI model before release. What risks are we looking for? How much does it slow things down? If the government takes months to greenlight something, meanwhile other countries have already shipped it. It's like asking someone to check every page of a book before publication—theoretically doable, practically a nightmare.

Worth noting: this is one of those rare moments where left and right agree AI needs oversight. The disagreement is on who oversees and how. Trump sees it as national security. The rest of the world (Europe included) is already moving in similar directions, just with different rules and less voluntary vibes.

The timing isn't random. As AI gets more powerful and widespread, whoever's in charge feels the need to have a hand on the lever. The problem is, pull it wrong and you might freeze everything.