We are the first generation that cannot trust its own eyes. A convincing video of a candidate saying something they never said — confessing to a crime, slurring a group, announcing they've dropped out — can now be made by anyone with a laptop, in an afternoon, for free. Drop it online 48 hours before an election, when there's no time to debunk it, and it can move votes on a lie that looks exactly like the truth. Every honest person across the political spectrum should find that terrifying.
Minnesota tried to do something about it. And in doing so it ran straight into the hardest problem in American constitutional law, which I'm going to lay out honestly rather than pretend it's simple, because it isn't.
What the law does
Minn. Stat. § 609.771 makes it a crime to disseminate a "deep fake" — defined as synthetic video, audio, or images "so realistic that a reasonable person would believe" they depict real speech or conduct that never happened — under specific conditions. To be criminal, the dissemination must be done knowing (or with reckless disregard) that it's a deepfake, without the depicted person's consent, with intent to injure a candidate or influence an election, and within a defined window: 90 days before a nominating convention, or after absentee voting has begun. Penalties climb with repetition, up to five years and a $10,000 fine for a repeat offender. It carves out broadcasters legally required to run candidate content.
Notice how narrowly it's drawn. It doesn't ban satire generally, or every altered image, or unflattering memes. It targets realistic fakes, made to deceive, aimed at an election, in the window where the damage is hardest to undo. The drafters clearly tried to thread the needle.
Then came the lawsuits
They threaded it, and got sued anyway — repeatedly, which tells you the question is genuinely hard. In Kohls v. Ellison (filed in federal court in Minnesota in 2024, now on appeal to the Eighth Circuit), a political satirist and content creator argue the law violates the First Amendment. Then X Corp. — Elon Musk's platform — filed its own challenge in 2025, and xAI, the artificial-intelligence company, filed another in 2026. When a satirist and two tech giants all sue the same short statute, it's a sign the constitutional problem is real, not manufactured.
Their argument deserves a fair hearing, because it isn't frivolous. The First Amendment protects a staggering amount of false and offensive speech — the Supreme Court has held that even outright lies are often protected, and that political speech gets the most protection of all. Satire and parody, which deliberately depict people saying things they never said, are core protected expression — that's what The Onion and every political cartoonist do. So the worry is: who decides what's a malicious "deep fake" versus protected satire, exaggeration, or commentary? A law that lets the government criminalize political content based on its judgment of the speaker's intent is exactly the kind of thing the First Amendment exists to distrust — even when the content is genuinely ugly.
Where I come down — and where I won't pretend
I'll tell you my honest instinct, and its limits. I think there's a real, defensible category here: a photorealistic fake of a specific candidate, made to deceive voters into believing they said something they didn't, dropped in the window where truth can't catch up. That's not satire anyone mistakes for commentary — it's fraud wearing a face. States have always been able to regulate fraud, defamation, and deception, and synthetic media is a new and dangerous delivery system for all three.
But I won't oversell it, because the line is genuinely hard to draw and the courts haven't finished drawing it. The danger of a law like this is that it's written or enforced too broadly — that it chills the cartoonist, the parody account, the heated-but-honest attack ad — and the remedy for that isn't to shrug at deepfakes, it's to draw the statute as narrowly as humanly possible and let the courts test it. That litigation is happening right now, in Minnesota, and it should. This is the system working: a legislature acts against a real threat, and the courts check whether it went too far. Anyone who tells you the answer is obvious — that deepfakes are simply "free speech," or simply "fraud" — is skipping the actual problem.
What we can do
Defend a narrow version, honestly. The state should defend the law as what it is — a tightly targeted fraud-and-deception rule for realistic fakes of real candidates near an election — and be willing to accept the courts' limits where the drafting reached too far. Overclaiming loses in court and deserves to.
Protect satire explicitly. Whatever survives should carve out parody, satire, and obvious commentary in the clearest possible terms. The cartoonist is not the enemy; the forger is.
Invest in the non-criminal defenses too. Fast, credible debunking, provenance and labeling standards for synthetic media, and media literacy do work a criminal statute can't, without the constitutional risk. The law is one tool, not the whole answer.
We can't trust our eyes anymore, and democracy runs on shared reality. Defending that is worth doing — carefully, narrowly, and with enough honesty to admit the First Amendment is watching, as it should be.
First the facts. Then the fix.
Sources
Minn. Stat. § 609.771 (Use of Deep Fake Technology to Influence Election), enacted 2023 (2023 Minn. Laws ch. 58; amended 2024 Minn. Laws ch. 112), verified against raw text at revisor.mn.gov: the definitions of "deep fake" (a realistic technological representation that a reasonable person would believe depicts real speech or conduct, substantially dependent on technical means) and "candidate" and "depicted individual" (subd. 1); the offense elements — knowing or reckless dissemination, without consent, with intent to injure a candidate or influence an election, within 90 days before a nominating convention or after absentee voting begins (subd. 2); the broadcaster carve-out; and the graduated penalties up to five years and a $10,000 fine for a repeat offense (subd. 3). Litigation verified on CourtListener: Kohls v. Ellison, No. 0:24-cv-03754 (D. Minn., filed Sept. 27, 2024), on appeal as Christopher Kohls v. Keith Ellison, No. 25-1300 (8th Cir., filed Feb. 12, 2025); X Corp. v. Ellison, No. 0:25-cv-01649 (D. Minn., filed Apr. 23, 2025); and X.AI LLC v. Ellison, No. 0:26-cv-03425 (D. Minn., filed July 27, 2026). The First Amendment principles (protection of false statements, political speech, and satire/parody) reflect well-established Supreme Court doctrine.
The current merits posture of each challenge (rulings, injunctions, appeals) was not read in full this pass and is described only as pending. Corrections: campaign@madgettformn.com.