Colorado deepfake ad of Sen. Cathy Kipp draws free-speech fight

The Independence Institute asked a Colorado administrative hearing officer on October 8 to dismiss the state's first campaign finance case over an undisclosed election deepfake, a fake video built with artificial intelligence that depicts state Sen. Cathy Kipp, a Fort Collins Democrat, working as a grocery clerk and keeping a customer's change.

The Elections Division of the Colorado Secretary of State filed the complaint on September 17 in case ED 2026-81, alleging the Denver think tank distributed, published and displayed a communication about an incumbent officeholder that contained an AI-generated deepfake without the clear and conspicuous disclaimer required by C.R.S. § 1-46-103. The division asks for penalties under § 1-45-111.5(1.5)(c.5). The General Assembly's summary of the act says a hearing officer may impose a civil penalty of at least $100 for a violation involving unpaid advertising, or at least 10% of the amount paid or spent to advertise the communication.

The statute, enacted as HB24-1147 and effective July 1, 2024, bars distributing a communication about a candidate or incumbent officeholder that includes a deepfake within 90 days of a general election unless it carries a disclosure stating that the material "has been edited and depicts speech or conduct that falsely appears to be authentic or truthful." Section 1-46-103(3)(e) exempts media content that "constitutes satire or parody."

Kipp filed the underlying complaint on August 5, arguing that the 45-second fake video, posted to the Institute's Facebook and X accounts, carried no compliant disclosure, Colorado Newsline reported. The division opened an investigation on August 19 and filed the formal administrative complaint on September 17, with the Attorney General's office as counsel of record. Colorado is one of 31 states that regulate deepfakes in election messaging, according to the National Conference of State Legislatures, as cited by Colorado Newsline.

That division complaint sets out the state's position: "satire and parody, at bottom, are grounded in humor." It argues the video is "express political advocacy, not satire" because it does not parody identifiable statements or actions by Kipp. The complaint notes that the clip's own on-screen notice, "AI generated content because who gets a human cashier these days?", ran in smaller type than the other text and appeared for about three seconds, against a law that requires the disclosure to be no smaller than the largest text in the communication.

David Lane of Killmer Lane, representing the Independence Institute, filed the motion to dismiss. He argues the satire exemption turns on whether a reasonable person would read the video as parody rather than a record of a real event, and that no reasonable person would believe Kipp works as a grocery clerk. The motion concedes the Institute's own on-screen notice "is not in complete compliance" with the font requirement, but argues the Institute had no obligation to include any notice if the satire exemption applies.

Lane told CBS News Colorado that the fake video would be a different matter if it showed Kipp making a racist speech or taking a bribe, and that he plans to sue Kipp and the state for attorney fees. He said the Tenth Circuit U.S. Court of Appeals has found that bad-faith investigations alone can violate the First Amendment, according to CBS Colorado.

Jon Caldara, president of the Independence Institute, told Colorado Politics the video is parody and that government actors are fair game. The Institute has opened a video contest paying $2,000 per ballot measure, Proposition NN among them, for AI-generated opposition videos; its published rules bar deepfakes of real people without consent.

Kipp has said the use of the fake clip is the issue. "The bottom line is simple," she told Colorado Politics. "It is disturbing and wrong to use a deepfake, AI-generated video of someone without their consent."

The case is the first in Colorado to test the 2024 disclosure law, and it lands amid a run of rulings against similar statutes. On September 18, U.S. District Judge Susan Watters granted a preliminary injunction against Montana's new deepfake law, finding it likely violates the First Amendment because it restricts deepfakes that criticize a candidate but not those that praise one, though she limited the relief to the named plaintiffs, the Daily Montanan reported. In the European Union, the AI Act's Article 50 labeling duty for deepfakes took effect August 2, 2026, requiring deployers to disclose deepfake content at first exposure.

For creators making political work, the practical question is where disclosure ends and the satire exemption begins. An administrative ruling that satire means any obviously staged clip would narrow the disclosure duty for AI political video. A ruling that it requires humor would keep most attack ads inside the rule. This case settles neither question for courts, because it is an administrative proceeding.

The hearing officer has set an evidentiary hearing for November 17, 2026 at 10 a.m. at the Secretary of State's office in Denver; the Independence Institute's answer is due October 19.