A federal judge last week granted a preliminary injunction against a new California law prohibiting the use of AI-generated deepfakes in political advertisements. This bill and two others like it have incurred the wrath of some on the right, including that of the Babylon Bee, the conservative satirical magazine.
The three bills in question were passed in response to a doctored video in which Vice President Kamala Harris referred to herself as “the ultimate diversity hire” and “a deep-state puppet.” Elon Musk, naturally, felt compelled to repost the deepfake, thereby blasting the disinformation to his 200 million followers. While some might be inclined to respond to this digital dumbshow with a sigh of “it is what it is,” California Gov. Gavin Newsom wasn’t among them. He and the Sacramento legislature, with its Democratic supermajorities in both houses, set about banning deepfakes in political advertisements.
The bills are complex, but taken together they (1) prohibit the use of deepfake technology to generate or modify the likeness or voice of elected officials, candidates, or election workers in “qualified political advertisements” during an election season; (2) require clear disclosure that AI technology has been used to generate or modify image, audio, or video, of anyone in a “qualified political advertisement”; and (3) create a system in which any “large online platform” (with more than 1 million users in California) is responsible for responding to complaints of undisclosed deepfake technology and removing such material if necessary.
This brings us to the strange part. The bills create a carve-out for “satire or parody.” Humorists can post deepfakes of political candidates, but must attach the following disclosure formula: “This ____ has been manipulated for purposes of satire or parody.”
This part of the bill irked the Babylon Bee, which filed suit against the Golden State (the injunction referenced above was in response to a different lawsuit). And what a suit it is. The Bee’s intervention reads like a cross between a senior undergraduate thesis, an op-ed, and a martyrdom video. The outlet even vows to continue breaking the law they are contesting, lawyerly norms be damned.
The Bee’s filing begins sensibly enough by declaring that satire’s goal “is often to criticize or mock an idea, event, or person for the purpose of correction and improvement.” But then the reasoning gets glitchy. For political satire to work, the Bee argues, audiences must not be initially aware that they are being exposed to political satire. California demanded that political deepfakes be accompanied by a disclosure. That seemingly sensible measure, the Bee alleges, would murder satire, ostensibly because satire is built on the obliviousness of those consuming it.
The power of satire, the Bee declaims, lies in its “proximity to the original.” Just like deepfakes, satire and parody “make audiences do a double take by believing that they are seeing a serious rendering of an original, and then allowing them to laugh at their own gullibility when they realize that they are really viewing satire or parody.”
Call it the abduction-van theory of political satire. Only if citizens are whisked away from their cognizance of what’s in front of them, blindfolded by wise comedians, and rolled up in a carpet—then and only then can satire fulfill its vital social function of “correction and improvement.”
True, some satire works by not letting audiences know that they are watching satire. But not all satire works that way. In fact, most satire does not work that way. When people tune in to The Daily Show or The Five, they know exactly what they are in for. Ditto for The Onion or the Babylon Bee. Yet that last outlet insists that satire only works when audiences are unaware of the satirist’s skillful manipulation. What kind of comedic power trip is that?
Is the Bee’s intervention indicative of a pro-deepfake column in the conservative movement? I don’t think so. Blue California wasn’t behaving that differently from red Florida, Tennessee and Oklahoma. Indeed, state houses nationwide are trying to regulate a technology whose lucrative, nihilistic credo seems to be “it is not what it is.”
If the Bee’s stance reflects a broader conservative trend, it has more to do with an embrace of maximalist conceptions of free speech than love of AI. The judge who granted the injunction reasoned that the free speech of the plaintiff (who posted the deepfake of Harris) was being “unconstitutionally chilled.”
Such a ruling is in line with a conservative dash away from a more limited understanding of the First Amendment (see, for example, the dissent of Justice Alito in U.S. v. Alvarez in 2012 and compare it to the conservative majority’s thinking about religious free speech in Kennedy v. Bremerton School District 2022). The ruling follows a judicial trend of prioritizing expressive liberty above anything else, like the need to not have American minds kidnapped en masse by disinformation.