The protests in Iran are real. The country’s economic desperation runs deep, and millions of citizens want to see a corrupt and repressive regime gone. The violent crackdown on the protests is also real and appears to have cost thousands of lives. Yet the accounts, photos, and videos coming out of Iran are riddled with accusations of AI manipulation and fakery that have the effect of calling even what’s true into doubt.
There’s a term of art for the benefit that accrues to bad actors when they sow public uncertainty as to whether anything at all is real: It’s the liar’s dividend, and it’s very much in evidence in Iran.
AI-generated or -enhanced online content has become a global problem for those seeking to understand or document protest movements, because it allows interested parties to shape narratives. In 2025, AI-generated images of real protest moments muddied the view of events in Turkey, and AI-generated content became a mobilization tool for Nepal’s Gen Z uprising. But Iran is perhaps the most fraught arena of all. Multiple political factions, foreign governments, and the Iranian regime itself are all competing to shape the narrative of current events. And the tools for fabricating reality have never been as advanced and accessible as they are today.
[Arash Azizi: The Islamic Republic will not last]
Videos of Iran circulating on social media show crowds moved by an array of convictions and desires. Some simply chant against the regime; others chant for Reza Pahlavi, the crown prince of Iran’s deposed monarchy, or invoke his father, the last shah. Numerous confirmed videos show demonstrators venting economic anguish and hatred for Iran’s supreme leader, Ayatollah Ali Khamenei; the security forces; and the Islamic Republic itself. All of these views of the protests are grounded in Iran’s complex reality. Yet the presence among them of images and videos that have been—or are even suspected of being—doctored with AI, mislabeled, misattributed, or spliced with content from other places or times has given oxygen to the dark imaginings of the movement’s enemies within the regime.
As a diaspora Iranian who has studied online deception, censorship, and surveillance efforts for almost 15 years, I take a particular interest in Iran’s information environment. At the human-rights organization Witness, I’m involved with something called the Deepfake Rapid Response Force, which helps journalists and human-rights defenders assess manipulated media. We have plenty to work with during the current Iran crisis.
What we are seeing is not a single disinformation campaign but something much messier. Iran’s information ecosystem has been shaped by extreme distrust of the Islamic Republic’s official media; social-media influence efforts coming from both the regime and foreign powers; and rampant distortions of online content—some deceptive, some careless, some well intentioned—that make verifying what Iranians are actually saying or doing during these protests difficult. The liar’s dividend redounds to the regime more than anyone else.
Iran’s nationwide uprising erupted on December 28. And within hours, regime-aligned accounts began dismissing authentic images from the protests as AI-generated.
An early, high-profile case involved a picture that came to be known as Iran’s “tank man” photo. Its source was a low-quality video that had been captured with a zoom lens on December 29, showing a protester facing down security forces in Tehran. A BBC Persian journalist posted a screengrab from the video the same day, and that image, which had a higher picture quality, began to circulate. The event was real: It has been verified from multiple angles and by fact-checkers. But somewhere along the line, someone enhanced the still image of it with AI editing tools, likely to make the blurry original screengrab sharper and more shareable.
Regime accounts seized on visible artifacts from AI enhancement as opportunities to dismiss the photo and other footage from the protests. “It’s all AI slop,” one regime account with a visibly AI-generated profile photo opined. “Just when you think Zionists can’t be more pathetic, they prove you wrong.”
Deepfakes have given AI a reputation for being deceptive. But many commonly employed image-editing tools use generative-AI capabilities, for better or worse. The public does not generally know how to distinguish good-faith use of these tools from malicious fakery. And this allows bad-faith actors such as the Islamic Republic, which has spent decades undermining protests and dissidence, to use not only AI itself but also the suspicion it generates as an accelerant.
The same pattern has played out with audio manipulations. On December 30, an account presenting itself as affiliated with the Mujahedin-e Khalq—a controversial exiled opposition group known as the MEK that has an armed wing and was once designated a terrorist organization by the United States—posted a video in which pro-Pahlavi chants are clearly dubbed over protest footage. The account then “exposed” the video as manipulated and used it to claim that monarchists were fabricating support. Within hours, pro-regime accounts had reposted this content to amplify the accusation that pro-Pahlavi chants were fake.
What was striking was that the alleged MEK account both introduced the manipulated video and was the first to call it out. Moreover, regime-aligned accounts were ready to rebroadcast that post almost immediately. Iranian researchers have long documented the regime’s use of opposition personas on social media to sow confusion, divide the opposition, and discredit authentic documentation, and leaked documents and videos from within the regime have confirmed the practice.
Within a day of the supposed MEK post being published, a Russian propaganda account set the AI-enhanced “tank man” photo alongside the allegedly manipulated Pahlavi audio and presented both as proof that the protests were a foreign fabrication. The loop was closed: manipulated content introduced by what may have been a regime account, amplified by pro-regime voices, then aggregated by Russian state-aligned media as evidence of an AI-driven Western psyop.
If that cycle isn’t complicated enough, consider that the regime’s accusations capitalize on at least one documented instance of manipulation by monarchists. On December 29, Manoto, the London-based Persian-language channel aligned with the monarchist opposition, aired footage of crowds chanting in favor of Pahlavi. The audio is real—but traceable to a memorial for Khosrow Alikhordi, a human-rights lawyer whose suspicious death last month drew thousands of mourners. It had been lifted and overlaid onto unrelated protest footage. Whether the doctored video was planted by hostile actors or intentionally created by someone within the royalist camp, the effect was the same: It handed malicious accounts exactly the precedent they needed.
Royalists are not the only anti-regime camp that has either generated or circulated fake content that serves their narrative. On December 30, Radio Farda and other accounts shared a 2020 video, as though it were footage of the current protests, of protesters chanting, “No to the oppressor, be it the shah or the leader.” BBC Verify’s Shayan Sardarizadeh flagged the misattribution.
The Islamic Republic has a habit of attributing authentic, homegrown dissent to Israeli and American conspiracies. This supposed foreign subversion then becomes a pretext for repression. The latest crisis is no different in that regard—but now foreign actors are helping boost the regime’s narrative by sharing doctored, misleading content that the regime can then expose.
Israel has been using AI-generated content to push anti-regime and pro-Israeli narratives in Iran since at least last summer’s 12-day war. The Citizen Lab, a research and policy center based at the University of Toronto, points to AI-fabricated footage of a supposedly surgical strike on Evin Prison, among other examples. A Haaretz investigation found that Israeli influence operations were promoting ideas aligned with Iranian monarchist factions.
Shortly before the protests broke out last month, an Israeli diplomat shared a montage of four alleged strikes on Iranian command centers during the June war captioned, “very precise.” Our analysis found that of the four clips, one is authentic footage that had aired on Iranian state television; the other three show signs of having been generated by AI. BBC Verify traced the video montage to an Instagram account that has more than 120,000 followers and identifies itself as “Islamists and Khomeinists’ Nightmare” and “Media Warfare Expert.” When users pointed out the manipulation of the air-strike footage, the diplomat did not delete the post. He acknowledged that parts of it may have been AI-manipulated, but argued that the video still demonstrates the precision of Israel’s operations—in other words, he stood by admittedly fabricated content because it illustrates a story he wanted to tell.
Perhaps similarly, on January 1, the Israeli Foreign Ministry posted an image on its Persian-language X account of Iranian police blasting two protesters with a water hose. The image was AI-altered: The original, published by BBC Persian’s Farzad Seifikaran the day before, shows the protesters being hosed down from a distance and then an officer approaching with a loudspeaker, not a hose.
These sorts of interventions may be meant to boost the opposition to the Islamic Republic, but the use of doctored media has other, no doubt unintended, effects. Actors with a vested interest in hiding what is happening in Iran’s streets use the existence of untrustworthy content promoted by foreign powers to convince the public that all documentation of the protests is a foreign deception. The dividend again accrues to the Islamic Republic.
Sometimes the problem is not manipulation at all but just the impossibility of capturing a fractured reality. On December 29, Iran International shared footage from a march in the town of Malard, where a crowd was chanting for Pahlavi. Another user posted a different angle of the same street and the same march, only with different audible slogans. One user who had shared the first footage issued a correction; others accused the Pahlavi camp of spreading falsehoods. But who is to say that on that long street, different crowds at different moments didn’t simply chant different things?
This is epistemic fog. Even without deception, people operating in good faith, trying to verify what they have seen, cannot know what is representative. Such knowledge is elusive in any case: Iran is a country of more than 90 million people whose aspirations can hardly be generalized from any single video or protest chant.
Journalists, courts, human-rights investigators, and protest movements everywhere have long depended on shared standards of proof. Today we are seeing what happens when those standards collapse. When any image can be dismissed as possibly synthetic, those in power no longer need to suppress speech. They have only to raise questions that undermine belief in a shared reality.
Since Thursday evening, the Iranian regime has thickened the fog to near impenetrability with a national internet shutdown and a blackout of communication with the outside world. From where I stand, the information picture has fractured further. Verified footage has slowed to a trickle; what comes out largely passes through Starlink terminals. Diaspora Iranians are frantic, unable to reach family members. Death-toll figures vary wildly, as documentation organizations struggle to access data: The range now sits between 12,000 and 20,000 dead, but the full scale is unknown.
[Read: How Trump could help the people of Iran]
Our Deepfake Rapid Response Force has been trying to help validate legitimate human-rights documentation that is sometimes falsely tagged as AI. An enormous effort is afoot among human-rights organizations to accurately count the dead. Meanwhile, regime accounts have been disseminating footage from counterprotests in favor of the Islamic Republic—clips that some opposition accounts suggest are AI-generated, and that official news media vehemently insist are real. Everyone is primed to distrust everything, and the shutdown ensures that nothing—not the scale of the protests, the severity of the crackdown, or the state of the regime—can be witnessed in real time.
The Iranian regime has spent more than four decades undermining the credibility of dissent. Now it has AI not just as a tool for creating fakes but as a rhetorical weapon. Every glitch, artifact, and even protective blur becomes evidence that nothing can be trusted. In a world where policy makers and platforms invest in AI detection and fact-checking, perhaps ordinary people would not be fighting through such darkness, or bad actors reaping such dividends.
What Iranians want and are fighting for is a question only they can answer. And their answer deserves to be heard—not shaped by foreign governments, dismissed by regime propagandists, or drowned in fog that others have helped create.
