Minnesota ban on election 'deepfakes' doesn't exempt parody, groups ask appeals court to clarify

Political satirists are ramping up memes and parodies with less than eight months until the midterm elections, emboldened by court rulings against bans on "materially deceptive" content, in the name of stopping "deepfakes," in California and Hawaii.

Minnesota could put a damper on the parody party, however, unless a federal appeals court reinstates a First Amendment challenge to its own deepfake law.

Satirist Christopher Kohls, his social media fan Minnesota state GOP Rep. Mary Franson and public interest law firms asked the full 11-judge 8th U.S. Circuit Court of Appeals to review a three-judge panel's ruling last month that Kohls, who earlier prevailed against California, was not actually at risk of prosecution because he labels his parodies as such by default.

This case "shows a trend toward narrowing state law to avoid a constitutional conflict, in this case by holding that the law does not criminalize constitutional parody despite no written parody exception," the Liberty Justice Center and Southeastern Legal Foundation wrote in a joint friend-of-the-court brief.

It is "inconsistent with the Supreme Court’s standing jurisprudence" and marks "a subtle departure from a textualist method of statutory interpretation" by letting courts rewrite constitutionally suspect laws "when considering standing in First Amendment preenforcement challenges," they wrote.

The panel also said Kohls, better known as "Mr Reagan," gave no evidence the law chilled others from sharing his best-known video, a fake campaign ad for Kamala Harris in the 2024 election, with an AI-generated voice, labeled as "parody" only in the title.

It drew more than 240,000 reposts on X, including by X owner Elon Musk and Franson. And Kohls did not show he lost income from the chilling effects of the law, the panel said. (It's not clear where the panel got the 240,000 figure. Kohl's own post shows 62,000 reposts and Musk's repost, another 227,000.)

X's own legal challenge, alleging the Minnesota law is preempted by Section 230 of the Communications Decency Act, has been stuck in a different court since December 2025, when a judge found X isn't likely to face enforcement and hence lacks legal standing.

While Franson's legal standing to sue was upheld, since she plausibly violated the law by not labeling her repost of Kohl's Harris parody as such, the panel denied her request for a preliminary injunction, saying she "unreasonably delayed in seeking relief," only challenging the law 16 months after voting for it as a state representative.

"The panel opinion adds to the growing confusion over important First Amendment and preliminary-injunction principles announced by" both the 8th Circuit and Supreme Court, the plaintiffs' lawyers at the Upper Midwest Law Center and Hamilton Lincoln Law Institute wrote in a joint petition for en banc, or full court, rehearing.

They cited a full 8th Circuit ruling in December that Missouri teachers had standing to challenge mandatory antiracism training in their school district, and Supreme Court 2020 ruling against New York's COVID-19 restrictions on religious gatherings. 

Courts supposed to avoid 'fuzzy question' of what law 'actually' prohibits

Unlike the California and Hawaii challenges, in which judges nominated by presidents George W. Bush and Joe Biden made quick work of the deepfake laws as First Amendment violations that broadly swept in parody, the Minnesota litigation has gotten hung up on technical questions about whom the law affects and evidence oddities.

Defendant state Attorney General Keith Ellison even got caught submitting an expert declaration with an "AI hallucination," a fake journal article. Stanford Social Media Lab founding director Jeff Hancock admitted to using ChatGPT but not carefully vetting its citations.

U.S. District Judge Laura Provinzino said she didn't consider either Hancock's stained or corrected declarations in rejecting the plaintiffs' motion for preliminary injunction, since she ruled based on lack of irreparable harm, not likelihood of success on the merits.

The cases by Kohls and Franson and Musk's X have received relatively little attention given the involvement of the world's richest man and the platform he controls, an influential source of campaign season parodies and memes. No one beyond the Liberty Justice Center and Southeastern Legal Foundation filed friend-of-the-court briefs.

The law unambiguously covers parody, the plaintiffs' petition for rehearing says. 

"To qualify as a deep fake under the statute, a video must be so realistic that a reasonable person would believe it depicts speech or conduct of an individual who did not in fact engage in such speech or conduct," the law reads.

"That’s it. There is no parody exception, no political-speech exception, no safe harbor for proper disclaimers, and no other exception that one may think of," the petition says.

The panel, appointed by presidents George H.W. Bush and George W. Bush, departed from 8th Circuit and SCOTUS precedent by tangling with the "fuzzy question" of what the law actually prohibits, when the standing inquiry is supposed to be "lenient," the plaintiffs said.

SCOTUS has repeatedly affirmed challengers don't have to say they will violate a law to challenge it, and that legal injuries can stem from both the deterrent effect of a law without "a direct prohibition against the exercise of First Amendment rights" and a prosecution that ends without a defendant's conviction, the petition says.

Some 8th Circuit panels started botching the correct standard "to our knowledge" in 2022 by replacing the SCOTUS qualifier "arguably" with "actually," which had real consequences in X's challenge to the deepfake law, according to the filing. Judge Provinzino admitted there she was "not entirely clear" whether the standard was "arguably" or "in fact" prohibited conduct.

The 2nd, 3rd, 5th, 9th and 10th circuits have all confirmed only "arguably" violating the law is enough to challenge it, with the 5th Circuit flatly declaring the "actually" standard "wrong," the petition says, quoting the reinstatement of a First Amendment challenge against the University of Texas for its speech code prohibiting "rude," "offensive" and "uncivil" speech.

"The panel resolved what is, at best, a difficult question of Minnesota statutory interpretation," yet Kohl's AI-generated Harris voice could easily qualify as a "technological representation of speech" that is "realistic" enough to confuse some people it's actually Harris, and the panel "seemed to agree" Franson posting the video itself violated the law.

By creating an "implicit safe harbor" for parody, the panel left Kohls and other speakers "with no way to know whether Minnesota could simply amend its statute" to eliminate that assumption "and then punish their speech," an obvious chilling effect, the petition says.

The public interest law firms said the full court should grant review "because the panel departed from textualist principles by either mistaking the context of the word 'realistic' in the Minnesota anti-deepfake law or silently applying the canon of constitutional avoidance to create a parody exception."

Textualism requires a word to be understood as would a person "conversant with our social linguistic conventions" and "from the context in which it is used," to quote the late Justice Antonin Scalia, and prohibits "what a judge thinks a legislature meant instead of what the words say," LJC and SLF wrote.

While the 8th Circuit has used constitutional avoidance when deciding between "competing plausible interpretations of a statutory text" in determining a case's merits, using the canon in the standing analysis "deprives a litigant of their day in court without the benefit of an opinion imposing that limiting construction on the statute," they said.

"By its ordinary meaning, in context, the law proscribes Christopher Kohls’ deepfake of Kamala Harris, with or without a parody label ... because its visual realism could fool people," in contrast to the legislation's exemption for "physical or verbal impersonation, the brief says.

The groups note Sen. Amy Klobuchar, D-Minn., called for X to remove the Harris video "because of her fear of voter confusion," undermining the panel's interpretation that applying a parody label to the video description "makes the speech unrealistic to the reasonable person" and hence doesn't violate the statute.