Anointed by AI: The Trump Pope Meme and the Expanding Liar’s Dividend

Belle Torek is an attorney who works at the intersection of free expression, civil rights, and online safety. This piece is submitted in her personal capacity and does not reflect the views of any current or former employer.

Just days after Pope Francis’ death, the official @POTUS and @WhiteHouse accounts published a joint post across social media platforms: an AI-generated image of President Donald Trump dressed in papal vestments and seated on the papal cathedra. The image followed Trump's quip to reporters that he’d “like to be pope” and would be his own “number one choice” as Francis’ successor.

The backlash was swift, especially among religious studies experts and Catholic communities in mourning, who called the image tasteless, even sacrilegious. The White House dismissed it as a joke, but in a subsequent press conference, Trump went further. “I had nothing to do with it,” he said. “Somebody made up a picture of me dressed like the pope, and they put it out on the internet. That’s not me that did it. I have no idea where it came from. Maybe it was AI, but I know nothing about it.”

Social media responses demonstrated immediate confusion as to what the president meant by these statements: was he merely acknowledging that the image was AI-generated? Or, more troublingly, was he alleging that AI itself posted the image to official government accounts?

Either interpretation is cause for concern, since both are attempts to wave away concerns from constituents to whom government actors should be accountable for official speech. Each version allows the president to shift blame from his administration to an amorphous technological scapegoat.

The first interpretation—that Trump was clarifying the image’s AI origins—seems both the least worrying and the least plausible. The image was clearly AI-generated (as are many others posted by this administration), making such a clarification unnecessary: still, it scapegoats technology and removes accountability from the government actors who posted it (and may have produced it). AI-generated images don’t appear ex nihilo—a human, likely one close to the administration, had to have prompted a system with something like “create Trump as the Pope” to generate it. Even if the image “maybe” came from AI, someone formed the query that brought it into existence, and someone posted it to official White House channels. And even if that person is not the president, they are part of his apparatus, duty-bound to answer to the public.

The second interpretation—that AI somehow autonomously posted the image—raises deeper questions. It signals an emerging reality where any digital act can be imputed to artificial intelligence, and even messaging from official government channels can be disavowed and blamed on machines. While it may seem preposterous, it’s not hard to imagine a politician with a loose commitment to the truth exploiting both the tools of generative AI and much of the public’s uneven understanding of how those tools work. The result is a potentially alarming expansion of the “liar’s dividend” with significant implications for accountability in government speech.

Coined in 2019 by legal scholars Robert Chesney and Danielle Citron, the “liar’s dividend” describes how deepfakes and synthetic media empower bad actors to deny real evidence: that wasn’t me; that’s a deepfake. Chesney and Citron predicted correctly that it would become a convenient exit ramp from accountability.

But this episode suggests a more evolved and insidious form of that evasion. When the liar’s dividend was conceived, the danger centered on authentic media being mistaken for convincing fakes. Here, by contrast, there’s little doubt the image was AI-generated. Under the second interpretation of Trump’s explanation, the deception would seem to lie less in questioning its authenticity than in obscuring its authorship and the process by which it was disseminated. In this case, the defense isn’t necessarily just That’s not real, but The tech made that on its own, or A human didn’t post that; the AI did. In such a scenario, the liar’s dividend has metastasized from a deceptive device to a rhetorical strategy, one that no longer disputes the truth of an image but instead seeks to dissolve human responsibility entirely.

The result is dangerous ambiguity: content that spreads with the imprimatur of government power but without clear speakers to hold accountable.

When government speech supposedly has no speaker

The Trump as Pope meme didn’t come from a personal account or even a parody account, but from official White House channels. A reasonable person would most likely understand it to be state, not private, speech: issued through the government’s accounts and cloaked in its institutional authority, regardless of who made and posted it.

Under the government speech doctrine, the state has broad discretion to express its own views–whether through a press release, a public health campaign, or, yes, even a social media meme. It can promote certain values, celebrate some causes over others, and express official stances. Critically, though, the constitutional check on government speech is not judicial review, but democratic accountability. If the public objects to state messaging, the recourse is the ballot box.

But that accountability mechanism only works if the public can identify who is speaking. As legal scholar Helen Norton emphasizes, democratic accountability depends on formal and functional transparency: the government should clearly claim its speech when authorizing it and signal its authorship when delivering it. If the source of a message is obscured, whether by design or dysfunction, then the public cannot meaningfully contest or critique it. In short, oversight evaporates when AI becomes a scapegoat.

We don’t know who created the papal meme image itself or who was behind the screen when it was posted. And as Charlie Warzel recently wrote in The Atlantic, we generally don’t know who runs the White House’s social media accounts or creates its memes. The memes may have originated from an official channel, but their authorship is entirely unknown. What we do know is that AI didn’t spontaneously generate or post the image on its own. Yet in a landscape where many Americans still lack a clear understanding of how AI works, the President’s vague, deflective language reads as an attempt to blur the lines and dodge accountability.

The situation evokes the meme of Spider-Man pointing at versions of himself: each account claims legitimacy, but none accepts responsibility. But unlike the cartoon, the stakes here are real because we lose the ability to check power effectively when we cannot find its source. The result is a constitutional hall of mirrors where state speech can be strategically disowned.

AI as both sovereign and scapegoat

This ambiguity lands in a moment of growing conversation about granting rights to AI. Some ethicists are now exploring whether chatbots should enjoy legal standing or even personhood. Public discourse has fixated on the prospect of AI sentience, even as these same systems hallucinate more often, regurgitate bias with ease, and become dangerously sycophantic in their responses to queries.

The Trump pope meme reflects the logical conclusion of that inversion: a world where machines are granted rights, and humans are excused from responsibility. AI becomes not just a tool, but a shield that protects powerful actors from the consequences of their own speech, and perhaps even conduct. This is no longer just a liar’s dividend, but a structural threat to the informational integrity that democracy requires.

Accountability–real, democratic accountability–requires attribution. The First Amendment enshrines the freedom to speak, but that protection becomes hollow when power speaks without consequence and hides behind a machine. In a democracy, leaders answer to the people. In the Catholic Church, the pope answers to God. In this new regime, AI answers to no one.