Colorado is the first in the U.S. to pass artificial intelligence regulations

The OpenAI logo is displayed on a mobile phone along with an image on a computer screen generated by ChatGPT’s Dall-E text-to-image model on December 8, 2023 in Boston. (AP Photo/Michael Dwyer, File)

Thanks to the stroke of Gov. Jared Polis’ reluctant pen, Colorado this month became the first U.S. state to pass a law explicitly regulating the use of artificial intelligence. It’s a milestone that supporters say provides an imperfect starting point for establishing oversight of the emerging industry.

The new state law broadly addresses the risk of discrimination when companies use AI while requiring a basic level of transparency.

“I believe Colorado’s AI law represents a major shift in the approach to AI oversight in the United States,” said Duane Pozza, a Washington-based attorney specializing in emerging technologies and artificial intelligence. “… This is the first law that tries to regulate artificial intelligence – or at least some types of artificial intelligence – more comprehensively and really puts a lot of requirements on so-called high-risk artificial intelligence systems.”

Starting in early 2026, the state will require certain companies that use artificial intelligence to make “consistent” decisions about disclosing the use and purpose of the technology to consumers, job applicants and others who interact with it. The law is intended to help Coloradans who can be vetted by an AI tool after applying for a job, financial service, education program or home or apartment.

By law, a job applicant vetted by an AI tool will be told they are dealing with a machine and explained why. If someone is rejected from a job or apartment, they will also receive an explanation.

Developers of such AI tools will also have to disclose more information about the systems, such as how they are tested for systematic errors. The measure is generally intended to limit the ability of artificial intelligence systems to discriminate against specific people or groups, supporters say, although it does not change existing discrimination law.

“Depending on how you program the decision-making process, (AI) can eliminate certain names or certain people based on race. There is inherent bias in some of these issues,” said state Sen. Robert Rodriguez, a Denver Democrat who sponsored the bill.

“All we are asking,” he added, is that companies “notify the person who comes into contact with them, and also conduct risk assessments and other activities just to demonstrate that they are updating the AI ​​when harm is detected. Because it will happen.”

Colorado’s law and similar attempts to pass legislation in other states have faced battles on multiple fronts, including between civil rights groups and the tech industry. Some lawmakers were wary of delving into technology that few yet understood, and governors worried about becoming weird and scary artificial intelligence startups. Polis expressed concerns about stifling AI innovation.

This year, states have debated many more bills targeting narrower snippets of artificial intelligence, such as the use of deepfakes in elections or the creation of pornographic videos.

When it comes to more comprehensive regulation, measures similar to Colorado’s have failed in states including Washington and Connecticut, whose proposal served as a model for Colorado’s original version. Another California bill has survived so far.

Colorado’s new law is likely one of the first broader artificial intelligence laws in the world, said Rep. Brianna Titone, an Arvada Democrat who co-sponsored the bill with Rodriguez and Democratic Rep. Manny Rutinel. The European Union also approved regulations on artificial intelligence this year.

Senate Bill 205 has been rewritten multiple times on its way through the state Capitol. Amid opposition from the tech industry, there were rumors that it would be vetoed – until the governor’s office announced that Polis had signed it in an announcement on the evening of Friday, May 17.

Still, he expressed doubts: In a statement accompanying his signature, the governor wrote that he was “concerned about the impact this bill could have” on the tech industry and artificial intelligence developers. He expressed hope that lawmakers would improve the measure before it fully enters into force on February 1, 2026.

Rodriguez, the Senate majority leader, said he didn’t think the bill would need further amendments next year, although both he and Titone said additional artificial intelligence regulation would undoubtedly be needed in the coming years.

This year, the legislature passed a companion bill that would expand an existing task force to study artificial intelligence.

“Extremely modest” requirements?

When the bill was introduced, it was greeted with skepticism from across the political spectrum. Dozens of companies, trade associations and employee groups have registered formal positions on the bill as the legislature debates it; no one said they supported it.

Labor groups were concerned that the proposed law could be used to circumvent existing anti-discrimination protections. Consumer advocates wanted stricter regulations. Technology and industry groups opposed it, calling it a stranglehold on a growing industry.

These groups argued that the federal government should take the initiative to ensure a unified approach.

After the amendments, labor groups and consumer advocates largely abandoned their open opposition and became more comfortable with the bill, although they were not very enthusiastic about it.

“It has some basic disclosure provisions that would at least shed some light on the murky world of AI-based decisions,” said Matt Scherer, senior policy advisor for labor rights and technology at the Center for Democracy and Technology in Washington, D.C.

“Besides, the requirements of the act are extremely modest,” he said. “And I would say that even the disclosure laws are far removed from what public interest groups are pushing for. But at least it’s a good reference point.”

Scherer said the bill includes exceptions for smaller businesses and companies to protect information they may consider a trade secret. He also argued that the law’s enforcement provisions should be strengthened, which would ultimately enable the Colorado attorney general to intervene against an infringing company.

Arguments against the state approach

Meanwhile, technology companies and industry groups remain opposed. Some opponents, including the national Consumer Technology Association, sent Polis letters asking him to veto the bill.

Doug Johnson, the association’s vice president of emerging technology policy, said in an interview Wednesday that he was concerned the bill would restrict emerging technologies and industries.