Two House lawmakers dropped a bipartisan discussion draft on Thursday that would create the first comprehensive national framework for artificial intelligence, temporarily blocking states from writing their own rules around AI model development. The three-year federal preemption window is designed to give the industry breathing room, but it also carries significant implications for the growing intersection of AI and crypto.
Rep. Lori Trahan, a Massachusetts Democrat, and Rep. Jay Obernolte, a California Republican, authored the draft. Their proposal would require leading AI developers to disclose the safety and security risks of their models under new federal mandates, while simultaneously preventing a patchwork of state-level regulations from fragmenting the industry.
What the draft actually does
The draft also includes provisions to expand AI research, signaling that lawmakers want to balance risk mitigation with innovation incentives. Top-tier AI developers, those building the most capable models, would face mandatory disclosure requirements around safety and security vulnerabilities.
This legislative push didn’t emerge from a vacuum. The White House released its own National Policy Framework on March 20, 2026, explicitly advocating for federal preemption of conflicting state laws. The Trahan-Obernolte draft is effectively the congressional companion piece to that executive vision, giving the administration’s goals legislative teeth.
Why crypto should be paying attention
The draft doesn’t mention crypto, blockchain, or decentralized protocols anywhere. Not once. But the second-order effects could matter quite a bit for a growing category of projects building at the intersection of AI and blockchain infrastructure.
The disclosure requirements are worth watching too. If federal rules mandate that top AI developers reveal safety risks, that standard could eventually flow downhill to decentralized AI platforms. The question becomes: who counts as a “leading AI developer” when the model is being trained across a distributed network with no single corporate entity at the helm?
The political landscape and what comes next
This isn’t the first attempt at federal AI preemption. Earlier efforts to slip preemption language into budget reconciliation bills met significant resistance from consumer protection advocates and state attorneys general who argued that removing state authority would leave consumers exposed. Those provisions didn’t advance.
The discussion draft format is deliberate. It’s not a formal bill yet, which means it’s designed to gather feedback before being introduced as legislation.
The bipartisan authorship is notable. Both parties broadly agree that the US needs to remain competitive with China on AI development, which creates unusual common ground.
For the Trump administration, this draft aligns neatly with its stated goal of making the US the global leader in AI. The March 2026 White House framework laid out the philosophical groundwork, and this congressional effort translates that philosophy into potential statute.