The most comprehensive AI regulation in the United States is effectively dead, and Elon Musk's xAI is the one holding the murder weapon.

On April 27, a federal magistrate judge issued an order freezing enforcement of Colorado's Anti-Discrimination in AI Law, which was supposed to take effect June 30. The law would have required companies using AI in hiring, promotions, and compensation decisions to implement risk management programs, conduct impact assessments, and disclose how their algorithms work. It was, by far, the most aggressive state-level AI regulation ever passed in the US.

Now it's frozen. And it might never come back in its current form.

How xAI Killed It

xAI filed suit on April 9 challenging the Colorado AI Act on constitutional grounds, arguing that forcing companies to disclose how their AI systems make decisions violates the First Amendment. The argument: AI model outputs are a form of protected speech, and the government can't compel companies to reveal their internal reasoning processes.

Then things escalated fast. On April 24, Trump's Department of Justice moved to intervene on xAI's side, raising additional constitutional challenges. The same day, Colorado's own Attorney General joined with xAI in a joint motion to temporarily stay enforcement. Read that again: the state's top law enforcement officer essentially agreed that his own state's law shouldn't be enforced.

The court's April 27 order prohibits enforcement until 14 days after it rules on xAI's preliminary injunction motion. That motion won't even be filed until 28 days after the state finalizes rulemaking. Translation: this law is on ice for months, possibly forever.

Colorado's Lawmakers See the Writing on the Wall

While the courts put the original law on hold, Colorado's legislature moved with unusual speed to replace it entirely. SB 26-189, introduced May 1, passed the Senate on May 7 and the House on May 9. Governor Polis is expected to sign it.

The replacement bill is dramatically narrower. Instead of the original law's "high-risk AI system" framework (which required impact assessments, risk management programs, and annual reviews), the new bill focuses on "automated decision-making technology" and mostly requires notice and transparency. Companies would need to tell people when AI influenced a decision about them and provide a plain-language explanation within 30 days of an adverse outcome. The compliance date is pushed to January 1, 2027.

Why This Matters Beyond Colorado

This is the first successful legal challenge to a state AI regulation law in the US, and it sets a dangerous precedent for anyone trying to regulate the industry. The First Amendment argument that xAI is making could apply to virtually any AI transparency requirement. If courts agree that AI outputs are protected speech and that disclosure requirements amount to compelled speech, the playbook for killing AI regulation just got a lot simpler.

And there's the irony of who's driving this: Elon Musk, who spent years warning that AI needed more regulation, now has his AI company leading the charge to dismantle the most significant AI regulation any US state has attempted. The DOJ intervention makes it even more pointed. The Trump administration isn't just watching AI regulation die. It's actively helping pull the plug.

Other states with AI regulation proposals on the books should be watching Colorado very carefully. The legal theory xAI is testing here won't stay in Colorado for long.