The pitfalls facing the bipartisan AI regulation bill

The draft artificial intelligence regulatory bill introduced by Reps. Jay Obernolte (R-CA) and Lori Trahan (D-MA) represents the best chance yet for significant federal action to address the fast-improving technology that threatens to disrupt society but has drawn opposition from groups that represent the industry, civil society, and labor.

The bipartisan discussion draft, called the Great American AI Act, would set a rulebook for AI companies, including standards for cybersecurity, the workforce, education, research, and international cooperation. It also addresses catastrophic risks and proposes mitigation measures.

Here’s what is in the bill, and some notable reactions.

A preemption of state laws

The legislation would impose a three-year ban on states implementing regulations on the development of AI models. Perhaps most notably, it would preempt a new California law requiring AI companies to disclose summaries of their training data, according to Trahan’s office.

Such federal preemption of state AI measures has been the subject of controversy over the past year. Many in the industry have favored a preemption measure to avoid complying with a complicated patchwork of regulations. Trump officials, too, have sought a preemption measure to facilitate innovation. But some grassroots activists, both left and right, have sought far-reaching state-level rules.

Rep. Chip Roy (R-TX), a staunch conservative, said this week that federal preemption is a “non-starter.”

Federal pre-emption on AI is a non-starter.

, Chip Roy (@chiproytx) June 11, 2026

Requirements for leading AI labs

The discussion draft proposes regulatory mandates for large “frontier” AI model developers, defined as those with more than $500 million in revenues.

These AI giants would be required to publish a “frontier AI framework” that discloses catastrophic risks, cybersecurity, and deployment decisions. They would have to publish safety reports before deploying new frontier models, and report any accidents to the Center for AI Standards and Innovation. Currently, the center exists as an entity within the Commerce Department created by Secretary Howard Lutnick. The legislation would codify it.

Frontier model developers would also be subject to audits by CAISI-licensed “independent verification organizations” aimed at identifying critical safety incidents. If the IVO discovers a potential hazard, frontier developers must report it. If they fail to do so or commit any other violation, the frontier developer will be subject to a civil penalty of $1 million per violation.

Those provisions have drawn some pushback from the industry. “The bill’s aggressive auditing regime and data-sharing requirements for developers could risk entrepreneurs’ trade secrets and private records,” Patrick Hedger, director of policy for tech industry group NetChoice, said in a statement.

Yet AI safety groups favor the measure. “The risk of not [relying on IVOs] is greater than not doing so,” said Brendan Steinhauser, the CEO of the AI safety watchdog Alliance for Secure AI Action.

The bill would also direct the Department of Homeland Security to provide coaching and protection to entities to help them mitigate AI threats.

J.B. Branch, the AI governance and technology policy counsel at the consumer group Public Citizen, said that the measures will fall short if they do not include a federal workforce for enforcement. “If you don’t have the federal workers to pursue this, it’s just words on paper,” he said.

Education

The bill would require the director of the National Science Foundation to implement strategies to incentivize research into equipping K-12 educators and students to address AI-related issues. They would also have to support scholarships and fellowships for undergraduate and graduate students pursuing AI-related degrees.

The bill also states that designated statistical agencies would have to collaborate to establish a “Workforce Research Hub” to evaluate the impact of AI on the workforce and generate actionable recommendations to mitigate unemployment.

The main concern

The biggest controversy, though, has surrounded the three-year preemption of state laws. In recent months, a number of states have moved to enact rules related to AI.

In New York, Democratic Gov. Kathy Hochul announced on June 9 that a new law to enhance AI transparency in advertising is now in effect. “In New York, we are setting the rules of the road instead of letting AI run the show,” Hochul said in a statement.

In California, Democratic Gov. Gavin Newsom signed an executive order last month requiring departments and agencies to incorporate findings from Engaged California’s current AI engagement into the executive order’s workstreams. Newsom has signed a number of AI measures into law beyond the safety disclosure measure, including ones related to chatbots and deepfake pornography.

Similarly, last month, Illinois lawmakers passed Senate Bill 315 to increase transparency and accountability in the largest AI models.

The Great American AI Act would not override those laws, but a number of interests have raised objections to the idea of one federal standard that preempts state action.

“We appreciate the attempt of this bipartisan bill that is taking the issue seriously,” Steinhauser said. “But we think the states should continue to lead on AI protections.”

Sacha Haworth, the executive director of the Tech Oversight Project, strongly opposes the draft bill. “The Obernolte-Trahan bill replaces a state floor with a federal ceiling,” Haworth said. “[It] trades away existing and future child safety, civil rights and consumer protection laws for nothing in return.”

She added that it gives big tech companies “the power to shape enforcement through the courts.”

Similarly, American Federation of Teachers President Randi Weingarten and Association of Flight Attendants-CWA President Sara Nelson issued a joint statement opposing the bill with a “hard no.”

“This bill is a giveaway to the AI industry and a handful of trillion-dollar companies, at the expense of American workers,” the statement says. “[It] blocks responsible state laws and gives free rein to a handful of billionaires who have already shown they cannot be trusted to put public interest ahead of profit.’

The discussion draft states that state attorneys general would still have a role in pursuing penalties against non-compliant AI companies. Still, the federal government can intervene at any time and immediately freeze and override any state-level lawsuit by filing its own federal case.

ELON MUSK BECOMES FIRST TRILLIONAIRE AFTER SPACEX’S BLOCKBUSTER IPO

Additionally, the bill allows states that voluntarily opt into the federal system to act as emergency co-enforcers or to opt out if they desire.

Reps. Scott Franklin (R-FL), Suhas Subramanyam (D-VA), Erin Houchin (R-IN), and Scott Peters (D-CA) support this bill and, like Obernolte and Trahan, hope to receive feedback from stakeholders, experts, and the public before the bill is formally introduced.