The Obernolte-Trahan AI Proposal: Five Positive, Important Takeaways
By Paul Steidler: This morning Representative Jay Obernolte (R-CA) and Representative Lori Trahan (D-MA) released a discussion draft, that is a framework for future Congressional legislation, for the “Great American Artificial Intelligence Act of 2026.” It is arguably the most important and consequential action Congress has taken on AI – ever – and provides a solid basis for future “must have” legislation.
Against this backdrop, and essential to understanding the need for legislation, is Mythos, the frontier AI model from Anthropic that the company said was too dangerous for public release and which it responsibly brought to the Trump Administration’s attention on April 7. This enabled banks and other critical infrastructure to take steps to protect against cyberattacks that could have been unleashed by the technology.
The potential havoc from Mythos struck fear deep in the hearts of many in the Trump Administration, though few may admit it. It led to a responsible and focused June 2 Executive Order from President Donald Trump that requests the federal government have 30 days advanced notice before such models are released, while also strengthening and better coordinating government cybersecurity programs.
Executive orders, though, can come and go with whoever is President. For the U.S. to have no policy for reviewing frontier AI models — that is, to wing it from administration to administration — raises the odds that harm will occur. Even more likely, and more ominously, it could put a chill on AI development by fueling more calls for an AI moratorium, and making such a moratorium more appealing to many.
The Obernolte-Trahan discussion draft aims to get it right regarding laws for reviewing frontier models, something which the federal government is uniquely qualified to do. The national interest will be served by preempting conflicting and burdensome state attempts to regulate this area.
The discussion draft also addresses workforce and cybersecurity issues, while leaving many other areas of potential AI legislation (e.g., energy use and copyrights) for discussion later and/or by others in Congress.
In sum, there are five points about the discussion draft that are especially compelling.
Focused. By focusing on frontier models, related cybersecurity, and workforce issues, it addresses the important AI issues today and those directly within the federal government’s purview.
Keeps attention on frontier models. Congress has a role — indeed, a duty — to understand the issues surrounding Mythos and other models with respect to cyber protection. The discussion draft and hearings on subsequent legislation will keep attention on these central issues.
Realistic. America will see other frontier AI model developments, and it will have leadership changes at the Presidency. Clear, compelling, targeted laws that stand the test of time are essential to govern how these models are reviewed.
Bipartisan. Republican Jay Obernolte led a bipartisan effort to release the U.S. House of Representatives Task Force Final Report on December 17, 2024 with broadly respected policy recommendations. Representative Trahan has also withstood political fire from those publicly chastising her for working with Representative Obernolte on the discussion draft.
Basis for changing the AI dialogue. The Obernolte-Trahan proposals should serve as the basis for high-profile Congressional hearings and related debate so the public can learn more about AI. Today, too much of the debate is driven by either those who want no regulation of the technology, or those who want to radically slow it down or even stop it altogether.
Neither approach is realistic nor do they pass the commonsense test. By getting real about AI, its proven ability to promote human flourishing while acknowledging risks and changes to employment need to be understood and discussed, America stays on the path of strong AI development, which will lift all of us.
Representatives Obernolte and Trahan have taken an important step with the discussion draft to help get us there.