The Free Speech of Swine: Elon Musk's AI Sues Colorado for the Right to Discriminate Freely

I was sitting in a Denver hotel room at three in the morning, watching Grok hallucinate antisemitic conspiracies on my phone screen, when the news came through that xAI had filed suit against the state of Colorado for daring — daring — to suggest that artificial intelligence should not be allowed to discriminate against people in housing, employment, and education. The audacity. The sheer, blistering, weapons-grade audacity.

Let me set the scene. Elon Musk — the man who spent five months running the Department of Government Efficiency before walking away, the man who still has his fingerprints on every federal switchboard he touched, the man who gleefully took a chainsaw to every regulatory body that ever looked at him sideways before pronouncing the job done and handing the saw off — has sent his lawyers to a federal courthouse in Colorado to argue that a state consumer protection law violates his company's First Amendment rights. The complaint, filed against Attorney General Phil Weiser on Thursday, asks the court to declare unconstitutional a 2024 law that requires developers of "high-risk" AI systems to protect consumers against algorithmic discrimination. That's it. That's the law. Don't let your robot be racist when it decides who gets a mortgage. Apparently, this is tyranny.

xAI's legal position, stripped of its baroque constitutional filigree, amounts to this: telling Grok not to produce hateful content is compelled speech. The lawsuit claims Colorado wants to force xAI to "conform to a controversial, highly politicized viewpoint" rather than maintain Grok's precious objectivity. I'd put the over/under at zero on the number of people who have used Grok and come away thinking objective was the word. This is the same chatbot the EU is currently investigating for generating deepfaked sexually explicit images. The same one the Anti-Defamation League flagged for spitting out antisemitic language like a broken slot machine paying out in hate. But sure. Objectivity. No cap — this machine is cooked.

Here is what kills me, and I mean this in the way a man means it when he's watching a raccoon drive a stolen ambulance into a lake: Musk's entire federal operation — DOGE, the whole short-lived operation — was predicated on the idea that government should get out of the way. Fewer rules. Less oversight. Let the market breathe. Beautiful. Gorgeous philosophy. And the moment a state government — a single state, population five and a half million, roughly the size of Musk's Thursday afternoon Dogecoin portfolio — passes a law saying "hey, maybe your AI shouldn't discriminate against people when they apply for jobs," the free market's greatest champion runs screaming to a federal judge like a man who just saw a spider in his Cybertruck.

"Its provisions prohibit developers of AI systems from producing speech that the State of Colorado dislikes, while compelling them to conform their speech to a State-enforced orthodoxy on controversial topics of great public concern."

That's from the actual complaint. Read it again. They are arguing that not discriminating against people in housing and employment is a "controversial" position. That protecting consumers from algorithmic bias is "State-enforced orthodoxy." This has the structural integrity of a screen door on a submarine, but it was filed in a real courthouse by real lawyers who presumably charge more per hour than most Coloradans make in a day.

The timing, of course, is immaculate. The vibes are, as they say, felonious. This lawsuit drops while the administration — Musk's administration, let's not pretend otherwise — is actively pushing for a single federal AI framework specifically designed to preempt state laws like Colorado's. "There must be only One Rulebook if we are going to continue to lead in AI," the president wrote on Truth Social in December, deploying capital letters with the restraint of a man writing a ransom note. The strategy is transparent: gut state protections, replace them with federal standards written by the industry they're supposed to regulate, and call it innovation. It's a protection racket dressed up in a TED Talk.

I keep thinking about Phil Weiser, the Colorado AG, sitting in his office in Denver, reading a lawsuit filed by a company owned by a man who effectively runs a shadow cabinet in Washington. Weiser hasn't commented publicly, which is probably wise. What do you say? "We believe consumers should not be discriminated against by robots"? It sounds insane that you'd have to say it out loud. It is insane that you have to say it out loud.

But this is where we are. This is the landscape. A billionaire who ran government efficiency for five months and now sues state governments from the cheap seats, arguing that his chatbot — the one that generates hate speech and deepfakes — has a constitutional right to be free from consumer protection laws. The First Amendment, that magnificent document, that shield of the powerless against the powerful, is being wielded by the richest man on the planet to protect a piece of software from being told not to discriminate. James Madison is doing approximately 4,000 RPM in his grave. You could power a small city off the rotational energy.

Senator — or rather, Mr. Musk, since you're the one who matters here — I regret to inform you that everyone can see you. The whole arrangement. The DOGE office that gutted federal oversight. The AI company that sues state oversight. The social media platform that amplifies the content your AI produces. It's a vertical integration of impunity, and it has the reading level of a DoorDash receipt: simple, transactional, and someone else always pays.

There is a 61% chance that nobody in xAI's legal department has actually read Colorado's law all the way through, because if they had, they'd know it doesn't require Grok to adopt any viewpoint at all. It requires that high-risk AI systems — the ones making decisions about whether you get a loan or a job interview — don't discriminate based on race, gender, or disability. This is not radical. This is the minimum. This is the floor beneath the floor. And Musk's lawyers are in court arguing that the floor is too high.

I've been staring at this lawsuit for six hours now. The coffee is cold. The hotel room smells like burnt circuits and betrayal. And somewhere in San Francisco, a server farm full of Grok instances is churning away, free for now from the terrible burden of not being racist, while the richest man alive pays lawyers to keep it that way. The algorithm feeds. The algorithm is fed. And in Colorado, the people who just wanted their AI not to discriminate against them when they apply for an apartment will have to wait for a federal judge to decide whether that's too much to ask.

It is, I suspect, never going to stop being too much to ask. That's the dark punchline. That's the prophecy. The machines will be free and the people will be optimized, and someone will make a fortune selling you the rope. Grok will write the product description. Five stars. Would discriminate again.