AI models face new government checks before public launch

The U. S. government just changed the rules for powerful AI tools. Instead of waiting for these systems to hit the market, officials now want a sneak peek—up to a month early. President Trump signed an order asking AI companies to hand over their most advanced models for review before wider release. The move targets systems that could impact cybersecurity, banking, or even military tech. Companies can still decide whether to join this review process. If they agree, the government will test the AI’s strengths and weaknesses behind closed doors. The goal? Spot any risks without forcing a full approval system. But critics wonder: could this slow down innovation, or is it just smart risk management? The order leaves room for debate.Why now? AI startups are racing to go public, with big names like OpenAI and Anthropic filing paperwork for stock market debuts. Some, like Anthropic’s new Claude Mythos model, focus heavily on cybersecurity—making them extra interesting to officials. But the Pentagon recently flagged Anthropic as a security risk, creating tension between progress and protection. The government’s approach isn’t about blocking AI—it’s about staying ahead. Federal agencies must draft new guidelines in the coming months, with a sharp eye on military and government systems. The big question remains: can regulators keep pace with fast-moving tech without holding it back?