White House Eyes Preapproval for Frontier AI Releases

The White House is considering a new executive order that would require government approval before companies can release powerful new AI models. This means advanced AI, like the kind from Anthropic's Project Glasswing, would need to pass tests before the public could use them. Here's the thing: officials have reportedly already blocked Anthropic from expanding access to its Mythos AI. The idea is to treat these AIs like new drugs, making sure they're safe before they go "into the wild." But, some critics worry this could lead to censorship or pressure on how AI models behave. Free-market groups and industry experts are pushing back. The bottom line: this move could either make AI safer for everyone or slow down innovation and limit who gets to decide what AI can do.