News & Publications
Election-related deepfakes are easy to create, harder to regulate
On Aug. 10, the FEC took the first step toward regulating federal candidates’ use of deepfakes and other AI-enabled fake images, audio, and video to depict and damage their opponents ahead of the 2024 presidential election.
Advancements in artificial intelligence have created a significant new concern ahead of the 2024 presidential election, and the Federal Election Commission (FEC) is grappling with how best to regulate this area and whether it can do so at all. Technological developments have made it possible to create realistic fake content at very little cost – experts warn the use of AI tools in the context of a fast-moving election could enable widespread misinformation and make it challenging for voters to distinguish what’s real from what’s fake.
On Aug. 10, the FEC took the first step toward regulating federal candidates’ use of deepfakes and other AI-enabled fake images, audio, and video to depict and damage their opponents ahead of the 2024 presidential election. By unanimous vote, the Commission authorized publication of a Notice of Availability of a rulemaking petition from Public Citizen, initiating a 60-day comment period to help the Commission decide whether to begin the rulemaking process.
Public Citizen’s petition specifically asks the FEC to clarify the scope of the fraudulent misrepresentation regulation at 11 C.F.R. § 110.16 to provide that a “deepfake audio clip or video by a candidate or their agent that purports to show an opponent saying or doing something they did not do” would violate the law.
The Commission, a six-member, evenly divided body that requires four votes to act, has already split on a threshold question – does the Federal Election Campaign Act of 1971 (FECA) provide authority to regulate deepfakes issued by a candidate? One commissioner (Allen Dickerson) has laid an early marker questioning the Commission’s power, stating that FECA only extends to “fraudulent agency,” meaning it covers “fraudulently pretending that you, yourself, represent or work for another candidate,” not “fraudulently claiming that your opponent did or said something they did not do.”
The statute in question provides: “No person who is a candidate for Federal office or an employee or agent of such a candidate shall fraudulently misrepresent himself or any committee or organization under his control as speaking or writing or otherwise acting for or on behalf of any other candidate or political party or employee or agent thereof on a matter which is damaging to such other candidate.” 52 U.S.C. § 30124(a)(1).
The skeptical commissioner has a point, at least to some extent. Congress has repeatedly rejected efforts to make the Commission a Ministry of Truth for campaign speech. For instance, amid debate on the legislative response to Watergate, Congress rejected an amendment from Senator Talmadge of Georgia to prohibit publishing “a false and defamatory statement about the character or professional ability of a candidate for Federal office.” 120 Cong. Rec. 9800 (April 4, 1974). Likewise, Congress rejected a proposal from Senator Tower of Texas to criminalize placement of “misleading advertisements” or “false oral or written statement[s]” about a candidate. 120 Cong. Rec. 8663 (Mar. 28, 1974).
Yet, the statute is broad enough to allow for regulatory clarifications. The legislative history shows that Congress sought to prohibit fraudulent conduct comparable to a candidate’s dissemination of a damaging deepfake depicting their opponent. Congress’ fundamental object was not addressing the technical issue of fraudulent agency. Rather, its purpose, in the words of Senator Birch Bayh, the statute’s sponsor, was to tackle Watergate’s “dirty tricks.” According to Senator Bayh, the statute was Congress’ “principal legislative response to the past 18 months of Watergate revelations” regarding fraud by the Committee To Re-Elect the President.
The dirty tricks that were discussed on the Senate floor included a 1972 incident where Nixon’s campaign forged a mailer on the letterhead of Democratic Senator Edward Muskie that accused his primary opponents (Senators Humphrey and Jackson) of “bizarre” conduct, falsely alleging that Senator Humphrey had a call girl in his car when he was arrested in 1967 for drunk driving, and that Senator Jackson had been arrested for homosexuality. 120 Cong. Rec. 10814 (Apr. 11, 1974); Martin Waldron, “Segretti Indicted in Mailing of Bogus Muskie Letter,” N.Y. Times (May 5, 1973). Thousands of copies of the unsigned letter were mailed to voters three days before the Florida primary. Id.
Importantly, the Bayh amendment struck a balance. It outlawed fraudulent conduct, without empowering the Commission to police campaign speech. Bayh informed his Senate colleagues that the proposal would “effectively deal with the specific campaign abuses which have been brought to our attention because of the 1972 campaign, without posing the difficult problems [of] a broader criminal libel statute,” like those proposed by Senators Talmadge and Tower. 120 Cong. Rec. 10814 (Apr. 11, 1974).
The distance between the fraudulent conduct proscribed by Senator Bayh’s amendment, part of federal law since 1974, and the conduct animating Public Citizen’s rulemaking petition is vanishingly small. Both efforts are designed to reach instances where a candidate issues a communication that puts words into the mouth of their opponent, thereby damaging their candidacy. It is not clear why the creation and dissemination of a mailer that falsely claims to have been written by an opponent would be prohibited by the statute (and subject to criminal penalties) while the Commission would lack authority to clarify by regulation that the creation and dissemination of a digital advertisement that falsely claims to consist of footage of an opponent is similarly prohibited. In neither instance would the Commission act as arbiter of truth; rather, the law would proscribe fraudulent conduct.
A carefully tailored regulation would likely survive judicial scrutiny. Numerous courts have held that the Commission has broad rulemaking authority and that its interpretation of FECA warrants substantial deference. These holdings stem from Congress’ grant of primary responsibility to the Commission to administer FECA and to “formulate general policy,” including through regulations that are “necessary to carry out the” Act, 52 U.S.C. §§ 30106(b)(1), 30118(a)(8); Doe v. FEC, 920 F.3d 886, 870 (D.C. Cir. 2019), and the Commission’s bipartisan structure, Teper v. Miller, 82 F.3d 989, 997 (11th Cir. 1996). In the landmark case Buckley v. Valeo, 424 U.S. 1, 110 (1976), the Supreme Court recognized that Congress gave the Commission “extensive rulemaking and adjudicative powers.” Given the Commission’s broad authority, courts will sustain an FEC regulation if it “is reasonably related to the purposes of the legislation.” Doe v. FEC, 920 F.3d at 252; see Orloski v. Federal Election Commission, 795 F.2d 156, 164 (D.C. Cir. 1986).
Given the close relationship between the purpose of the misrepresentation statute and the conduct sought to be proscribed by the rulemaking petition, the Commission has the authority to make policy in this area if it can muster the political will to do so.