(Illustration via Pixabay by geralt)

WHEN BROOKS BROTHERS shut down its longtime clothing factory in Haverhill in 2020, it didn’t just close a building. It ended careers that workers had spent decades building and sent a familiar shudder through Merrimack Valley communities like Lowell and Lawrence that know all too well what boarded up mill buildings leave behind.

The company cited costs. To them, the math was simple, even if here at home, we knew the human toll was not.

After decades of offshoring and automation hollowing out American manufacturing, the workers left behind in Haverhill didn’t need a study to tell them what was coming. They needed Congress to act years before they got handed pink slips. On artificial intelligence, we still have a chance to move before similar damage is done, but that window is closing faster than most people realize.

Earlier this year, Anthropic unveiled Mythos, an AI model capable of identifying thousands of vulnerabilities across every major operating system and web browser. It was deemed too dangerous for public release, but that has not stopped competitors from racing to catch up, or our adversaries from trying to access it. The speed of AI development means Mythos represents a floor on what these systems can do. The models will only get more powerful, and the risks to America’s workers, national security, and cyber infrastructure will only grow.

Despite these flashing warning lights, there is no federal law on the books governing how the most powerful AI systems in the world are built, tested, or deployed. No independent auditors verify the safety claims of the largest AI companies, commonly referred to as “frontier” labs. No federal agency has clear authority to step in when something goes wrong. While some have argued there is still plenty of time for Congress to act, I would say, look around.

New and recent college graduates are struggling to find work in their fields, facing higher unemployment rates than the broader workforce. Across the economy, workers are asking the same hard questions. Will the job they do today exist in five years? Will the prosperity this technology generates reach them or bypass them entirely, just like the promises of globalization did?

At the same time, our adversaries are racing to build and deploy AI systems capable of undermining our military advantage, penetrating our critical infrastructure, and eroding the democratic institutions that define us.

After Mythos, I went to work trying to change that. I entered negotiations with my Republican colleague on the House Energy and Commerce Committee, Congressman Jay Obernolte of California, to build a bipartisan federal AI framework. Those talks have been hard and remain unfinished, but they have made something clear: The basic architecture of serious AI governance is within reach. There is more bipartisan agreement on the fundamentals than people might expect.

Here is where that consensus lives.

First, real accountability at the frontier. The largest AI companies should be required to publish and comply with safety and security frameworks. They should have to submit to third-party audits to show their work, and federal and state regulators should be empowered to enforce and update those requirements as the technology evolves. The most powerful labs also should not be allowed to silence whistleblowers who want to expose wrongdoing. The companies building models that could reshape our future should not be operating on the honor system.

Second, independent verification of safety practices. The federal government cannot verify every AI model itself, nor should it try. Instead, it can accredit private organizations to embed within frontier labs, assess their safety practices, and call in federal and state enforcers when the companies fall short. These organizations must be nimble, transparent, and built for the speed of the technology while ensuring Congress, regulators, and the public know whether safety commitments made in press releases are being honored in practice.

Third, protect American workers. The lesson from Haverhill is that job training after the fact is not a policy. What workers need is a real-time picture of how AI is reshaping the labor market so Congress can get ahead of disruptions rather than respond years too late.

When AI is the reason workers get pink slips, employers should have to say so. Updating WARN Act requirements would mandate disclosures when AI drives a mass layoff, so we stop pretending that decisions like the one Brooks Brothers made happen in a vacuum. We need a framework built on the belief that the workers who built this economy deserve to lead in the next one.

Finally, shore up our cyber defenses. AI is only as secure as the open-source software it is built on, and right now the developers maintaining that code are largely on their own. Funding should be allocated for open-source maintainers, and frontier AI developers should be required to give them access to their models for security work. And the information sharing framework that lets companies flag threats to each other without antitrust risk must be reauthorized so a threat one company sees today cannot become a breach across the entire sector tomorrow.

Since entering these negotiations, I have heard many arguments about why Congress should wait to act. Some say the market will sort it out. We heard that about financial derivatives and social media, and the American people are still paying the price. Others counter that states are doing good work and the federal government should stand back. They are right that states have shown real leadership, but the best of that work can and should inform a federal framework.

Serious governance concentrates oversight at the frontier. It elevates the strongest provisions from the states – our laboratories of democracy – to create a federal standard, focused on model development, while explicitly preserving states’ ability to protect residents from AI systems that discriminate, deceive, or endanger. The Democratic leaders on Beacon Hill who have reached out to me about this work understand what is at stake for Massachusetts residents, and this framework is designed with them in mind.

I cannot go back to Haverhill, look a worker, a parent, or a recent graduate in the eyes, and tell them that we saw what was coming and decided to wait for a more opportune moment. Congress must act for the workers, the families, and the communities that cannot afford for us to get this wrong. Again.

Lori Trahan is a Democrat representing Massachusetts’s Third Congressional District. She is a member of the House Energy and Commerce Committee.