The Trump administration is closing in on a new federal framework for artificial intelligence, and one of the biggest unresolved questions is how Washington will treat open-source models. That’s according to The Information, which reports the policy is nearing completion even as the open-source debate remains unsettled.
This matters because it signals the government is moving from rhetoric to rules. For most of the past two years, US AI policy lived in executive orders, voluntary commitments, and agency guidance. A formal framework changes the game for anyone building, deploying, or investing in AI.
What we know
The core of the report is straightforward, and I’ll stick to what The Information has confirmed:
- The administration is close to finalizing an AI framework.
- Open-source questions are still open, meaning the government hasn’t settled how freely available model weights will be handled.
- The timing suggests an announcement could come sooner rather than later.
The open-source piece is the part to watch. “Open-source” in AI usually means open-weight models, where a company releases the trained parameters so anyone can download, run, and modify them. Meta’s Llama family, plus models from Mistral and others, made this approach mainstream. The upside is speed, transparency, and access for smaller players. The worry, voiced by some in national security circles, is that powerful open weights can’t be recalled once they’re out.
Why this is a turning point
The status quo before this was a patchwork. The previous administration leaned on a sweeping executive order and reporting requirements for the largest models. The Trump administration moved early to roll back parts of that approach and signaled it wanted a lighter, more growth-focused stance. A dedicated framework is the next logical step.
What stands out here is the tension baked into the process. A pro-innovation administration has every reason to protect open-source development, which American startups and researchers rely on. But the same administration has to weigh national security arguments about frontier models landing in the wrong hands. How it resolves that tension will tell you a lot about where US AI policy is heading.
What’s at stake for practitioners
If you build on open models, the framework could shape your roadmap in concrete ways:
- Access: Rules could affect which model weights stay freely downloadable and which face new conditions.
- Compliance: A formal framework may bring reporting or safety obligations for developers of the largest systems.
- Competition: Clear rules can favor big labs with legal teams, or protect the open ecosystem that keeps smaller players in the race.
- Global positioning: US choices ripple outward, especially as Europe enforces its AI Act and China pushes its own open models.
What comes next
Watch for the actual text. Frameworks live and die in the details, and the open-source decision is the one that will draw the loudest reactions from both the AI industry and the security community. Expect labs, open-source advocates, and civil society groups to push hard in whatever comment or rollout window follows.
For now, the signal is clear: formal US AI rules are close, and the fate of open-source models is the fault line. I’ll be tracking how the administration lands that call. Full reporting is available at the original source, The Information.