Colorado fights to shield residents from AI bias in hiring, housing, and healthcare

Colorado wrote the rulebook on AI decisions

Colorado now has the first broad AI law in the country, and it takes effect June 30, 2026.

Gov. Jared Polis signed SB 24-205 in May 2024, targeting AI systems that make or heavily influence major decisions about people’s lives. Think hiring, housing, healthcare, lending, and more.

Polis signed it “with reservations,” asking lawmakers to tighten it up before it kicked in. Other states passed narrower AI rules, but Colorado’s law covers the most ground.

The law targets decisions that really matter

The law zeroes in on what it calls “consequential decisions,” ones that carry real weight in people’s lives.

That means AI systems used in hiring, rental applications, healthcare access, education, lending, insurance, government services, and legal services all fall under the rules.

A decision counts if it has a meaningful legal or similar effect on someone’s access to those areas. The law does not cover chatbots or content generators.

It focuses on systems that make calls about people.

AI bias already shows up in everyday life

AI systems can bake in discrimination when they learn from data shaped by old patterns of bias.

Resume screening tools have filtered out qualified job applicants based on things like name or zip code that track with race or gender.

In housing, algorithms have set rental prices or screened tenants in ways that hit certain groups harder.

Lending systems have denied loans or charged higher rates using factors that stand in for protected characteristics. Colorado’s law defines all of this as “algorithmic discrimination.”

Companies face new rules on both sides

Starting June 30, 2026, companies that build AI systems and companies that use them must take “reasonable care” to protect people from discrimination.

Companies using these systems must run impact assessments before deployment, update them yearly, and redo them within 90 days of any big change.

They also need a risk management program tied to standards like the NIST AI Risk Management Framework. Builders must hand over detailed documentation on how the system works and what risks it carries.

Consumers get a say in AI decisions

Companies must tell people before an AI system makes or shapes a major decision about them.

If that decision goes against someone, the company has to explain why, describe how much the AI played a role, and say what data it used. People can correct the wrong personal data that the system relies on.

They can also appeal the decision and ask for a human to review it, if technically possible. Colorado’s existing privacy law also lets people opt out of AI profiling altogether.

Fines can stack up to millions fast

Violations fall under Colorado’s Consumer Protection Act as deceptive trade practices. The maximum fine is about $20,000 per violation, and each affected person or transaction counts separately.

So a system that discriminates against just 50 people could trigger up to $1 million in penalties. When the affected person is elderly, fines can reach about $50,000 each.

Only the Colorado Attorney General can enforce the law.

No private lawsuits are allowed, and the Attorney General must give 60 days’ notice to fix the problem first.

Smaller businesses catch a break

Companies with fewer than 50 full-time employees get some relief from the heaviest requirements.

They do not have to build their own risk management program or run independent impact assessments, as long as they do not train the AI with their own data and use it as the developer intended.

But small businesses still have to share the developer’s impact assessment with consumers and notify people when AI shapes a decision about them.

The law covers any business operating in Colorado or affecting its residents, no matter where the company is based.

Lawmakers already delayed the law once

The law was supposed to take effect Feb. 1, 2026. But Gov. Polis called a special session in August 2025 after businesses raised concerns about compliance costs.

More than 150 lobbyists showed up at the state capitol as four competing bills tried to change or kill the law. None of them passed.

The only result was SB 25B-004, which Polis signed on Aug. 28, 2025. It pushed the start date to June 30, 2026, without changing a single requirement.

The 2026 session has not touched it

Colorado’s regular legislative session started Jan. 14, 2026, and runs through May 13. Many expected lawmakers to water down or strip the law’s requirements.

So far, no deal has come together. Politics may play a role: Colorado elects a new governor in November 2026, and the current Attorney General, who would enforce the law, is running for the job.

If the session ends without changes, every original provision kicks in as written on June 30, 2026.

The White House called out Colorado by name

On Dec. 11, 2025, President Trump signed an executive order that singled out Colorado’s AI Act.

The order said laws banning algorithmic discrimination could force AI systems to produce false results just to avoid treating people differently.

It created a DOJ AI Litigation Task Force to challenge state AI laws that the administration considers inconsistent with national policy.

The Department of Commerce must also flag state AI laws it views as unconstitutional or overly burdensome, with a report due by spring 2026.

Legal experts doubt the order will stick

Executive orders are not laws and cannot override state statutes without Congress stepping in. Legal analysts at several major firms have said the order is unlikely to preempt Colorado’s law.

A proposed 10-year federal ban on state AI laws failed in the Senate in 2025 by a near-unanimous vote. Colorado’s Attorney General has pushed back publicly and called on Congress to take up AI regulation itself.

Governors in Colorado, California, and New York have all said the order will not stop their enforcement.

Other states are writing their own AI rules

Colorado is not alone. Illinois put an AI disclosure law for employers into effect on Jan. 1, 2026.

California finalized rules governing how employers use AI under its civil rights and privacy laws. Georgia and Texas have introduced or passed their own AI-related bills.

States are moving because Congress has not passed a broad federal AI law. What happens with Colorado’s law could shape how other states handle AI oversight going forward.

This article was created with AI assistance and human editing.

Read more from this brand:

• Colorado’s AI Law Bans Discrimination. The White House Calls It Unconstitutional.
• “Unsinkable” Molly Brown launched America’s first female Senate bid from this Colorado mansion
• What’s beneath Denver’s Union Station? A WWII time capsule few ever see