Netanyahu Posts Proof-of-Life Video as AI Sows Doubt About What Is Real
Misinformation has always been a serious problem, but artificial intelligence has taken the game to a whole new level. Think about it: a sitting prime minister having to record a video in a coffee shop, holding up his own hands for the world to see, just to prove he is alive. That is exactly what happened with Benjamin Netanyahu in March 2026, an episode that reveals how the era of generative AI is radically transforming the relationship between truth, imagery, and public trust.
After posting a video address directed at the Israeli people on social media on a Friday, the prime minister was bombarded with claims that the recording was fake — generated by AI. The discrediting campaign came mainly from online accounts with ties to Iran, which quickly turned the video into the target of a coordinated disinformation operation.
The main argument? That he appeared to have six fingers on his hands, that classic error that AI image generators used to make all the time. Many users claimed to have spotted a sixth finger, and the allegation went viral at an impressive speed. Even though fact-checking organizations like Snopes and Newsweek debunked the story quickly and confirmed the video was authentic, the narrative kept spreading. A single post on X surpassed 2 million views. 😬
This episode is more than a quirky story of the week. It raises an alarm about something that goes well beyond a viral video. AI is not just creating fake content — it is also being used as an excuse to discredit what is completely real. And this phenomenon has a very specific technical name: the liar’s dividend.
The Coffee Shop Video and the New Proof of Life in the Digital Age
Recognizing both the power and the absurdity of the situation, Netanyahu decided to respond in a way no world leader should ever need to. Two days after the original address, he recorded a new video, this time in a much more casual setting — a coffee shop in Israel. In the polished clip, he raised his hands in front of the camera and showed all five fingers on each one, in a gesture that became a kind of proof of life adapted for the age of artificial intelligence.
The scene became a symbol of a very specific moment in the history of digital communication. We have reached a point where distrust of online content has hit such an absurd extreme that even direct, informal visual proof can be called into question. The coffee shop where the recording took place even posted its own photos of the prime minister’s visit on Instagram — blurrier, less posed images that served as yet another layer of evidence about the authenticity of the encounter.
What fueled the original narrative was a technical detail a lot of people are already familiar with. For a long time, AI image generation models had serious trouble rendering hands correctly. Extra fingers, impossible angles, weird proportions — these were frequent and easily identifiable errors for anyone even slightly familiar with the subject. This knowledge, which was basically a technical data point about AI limitations, was turned into a narrative weapon by those looking to discredit Netanyahu’s original video. Accounts with alleged ties to Iran amplified the story across social media, and content distribution algorithms did the rest of the heavy lifting.
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
The most revealing part of all this was not the video itself but how quickly the lie spread even after it was debunked. Fact-checking organizations acted fast, analyzed the material, confirmed it was authentic, and published their findings. Even so, the narrative kept circulating and gaining traction. This shows that in today’s digital environment, the correction rarely reaches the same audience the original lie did — a structural problem that platforms still have not managed to solve.
The Liar’s Dividend: When the Existence of AI Becomes the Excuse
The term liar’s dividend, described by researchers in publications like the California Law Review, was coined to describe a dangerous side effect of the popularization of generative artificial intelligence. The logic is straightforward and alarming: the more people know it is possible to create fake videos, photos, and audio with AI, the easier it becomes for someone to claim that a completely real piece of content is a digital fabrication. In other words, the mere existence of the technology is enough to cast doubt on anything. And that doubt, even without any real basis, carries political, social, and strategic value for anyone looking to manipulate public opinion.
Alberto Fittarelli, a senior disinformation researcher at the Citizen Lab at the University of Toronto, was blunt in his assessment of the situation. He stated that anyone with knowledge of manipulation techniques and the willingness to use them would exploit the liar’s dividend to sow distrust about the realities of an armed conflict. Fittarelli also highlighted something fundamental: verifying everything is incredibly exhausting, and not everyone can afford that luxury.
In Netanyahu’s case, the mechanism played out in an almost textbook fashion. The original video was real. The claim that it was AI-generated was false. But the simple fact that the technical possibility of creating such a video existed was enough for many people to doubt it. The narrative did not need to be true to be effective — it just needed to plant the doubt. And doubt, once planted, is hard to remove, especially when the people who spread it have an interest in keeping it alive and continue amplifying the original content even after the fact-checks are published.
This phenomenon represents a structural shift in how disinformation operates. Before, the challenge was fighting fake content that looked real. Now, the challenge also includes defending real content that has been labeled fake. These are two opposite problems, but they share the same root: the erosion of trust in what we see and hear online. And the more sophisticated AI becomes, the harder it gets for the average user to tell one from the other without specialized tools or verified sources.
The Iran Conflict as a Disinformation Laboratory for AI
The Netanyahu episode did not happen in a vacuum. It sits within a much larger context: the war in Iran, which since February 2026 has produced thousands of images and videos — many real and many generated by artificial intelligence, with differences that are increasingly difficult to spot with the naked eye by anyone simply scrolling through their social media feed.
The liar’s dividend had already been playing a significant role before the armed conflict even began, during waves of protests against the Iranian theocratic government. A video confirmed as authentic by The New York Times showed a protester sitting peacefully in the street while heavily armed police advanced on him. The scene evoked the iconic Tank Man of Tiananmen Square in 1989 and spread widely online. Even with confirmation from multiple angles, pro-government voices dismissed the footage as an AI fabrication.
These claims, although false, gained traction because AI-generated content was in fact circulating at the same time. An investigation by the Citizen Lab, published in October 2025, revealed that the Israeli government or a subcontractor had used AI-generated content to encourage Iranians to overthrow their government. As Fittarelli noted, the fact that Benjamin Netanyahu had to prove he is alive and that his image was not generated by artificial intelligence shows that the risk cuts both ways.
With the start of the war in February 2026, videos and images captured the destruction in Tehran and across the country as bombings hit various targets. After a missile strike destroyed a girls’ school, killing at least 175 people in what a preliminary military investigation classified as an apparent targeting error by the United States, authentic videos of the destruction began circulating online. Some social media users, however, incorrectly claimed that the scenes of rubble, grieving parents, and mass graves were fake — created by AI or recycled from conflicts that had occurred years earlier.
Manipulation From Both Sides of the Conflict
Mahsa Alimardani, associate director of technology at Witness, a human rights organization that studies the impact of AI on video evidence, highlighted a particularly cruel aspect of this dynamic. She observed that the Iranian government had cut off the internet and tried to block documentation of protester deaths in January, but was now invested in detailing the fatalities linked to Israeli and American strikes. According to Alimardani, the regime is engineering the information environment, and at the same time it sowed the doubt that is now being weaponized against authentic documentation.
The Iranian government itself leaned into AI manipulation, circulating synthetic images to emphasize the high cost of the war. Alimardani pointed to the example of an image of a bloodied, dusty children’s backpack published by the Iranian Embassy in Austria. Although the image looked real, it had been created by Google’s image generator, as identified by the company’s own detector. This kind of use shows how the technology serves as a tool both for fabricating narratives and for undermining the credibility of real facts. 📲
The Role of Platforms and the Limits of Verification
Digital platforms play a central role in this equation, and honestly, they still have not found a satisfactory answer to the problem. Meta’s Oversight Board, the company’s quasi-independent body, issued a statement the week following the episode acknowledging that the problem existed during global conflicts and crises, including the situation in Iran. The board called on the social media giant to do more to identify misleading AI-generated content circulating during armed conflicts.
On X, where Netanyahu’s video racked up millions of views, the situation got even more complicated. Iranian state-affiliated media outlets, like the Tasnim News Agency, continued fueling the skepticism. Social media users claimed the coffee shop video had been derived from a photo taken in 2024 — even though the coffee shop in question only opened in the summer of 2025. Others posted similar videos from the same location to demonstrate how easy it would supposedly be for AI to generate similar clips, including versions with Netanyahu wearing a sports jersey and even with the new leader of Iran and other heads of state replicating his movements.
The episode took on even more troubling dimensions when Grok, the AI chatbot created by Elon Musk’s xAI, incorrectly backed the false claims right there on X. The chatbot wrote that Netanyahu’s video had been generated by artificial intelligence, calling it a classic deepfake meme — a post seen by more than 100,000 people. The company did not respond to requests for comment. The fact that an AI tool reinforced a false narrative about the use of AI adds a deeply ironic layer to the whole story.
Trust in Crisis: What Is at Stake Beyond a Single Video
The Netanyahu episode is not an isolated case. It is part of a growing pattern where political leaders, public figures, and institutions have to contend with claims that their communications are digitally fabricated. This dynamic erodes something far more valuable than any single person’s reputation: it erodes collective trust in the media, in digital platforms, and even in the idea that something like a verifiable fact exists. When anything can be questioned with the argument that it was made by AI, public discourse loses a fundamental foundation.
Detection technology for AI-generated content has also advanced significantly, but it is still far from foolproof. Tools like deepfake detectors, invisible digital watermarks, and media provenance authentication systems are being developed and refined by companies like Google, Adobe, and various specialized startups. Fact-checkers have also relied on supplementary evidence — images from other angles, records from journalistic sources, metadata — to determine the authenticity of materials. But these resources usually arrive after the damage is done, and they require a level of digital literacy that a large portion of the population still does not have.
In the meantime, disinformation continues to operate at the speed of social media, and trust keeps getting chipped away, one video at a time.
What This Moment Says About the Future of Information
The Netanyahu video case serves as a barometer for where we are right now. Generative AI has democratized content creation in a way that would have been unthinkable ten years ago, and that has brought real benefits for creators, businesses, and everyday users. But it has also introduced a new layer of complexity to the relationship between people and the information they consume. Today, seeing is no longer believing. And this paradigm shift has consequences that extend far beyond politics or technology — it affects how societies make collective decisions, how they choose their leaders, and how they build consensus about what is true.
Researchers who have studied disinformation for years warn that the problem will not be solved with better technology or more responsible platforms alone. A significant part of the solution lies in media literacy — in people’s ability to recognize manipulation patterns, question sources, and understand how algorithms amplify certain types of content at the expense of others. That is not simple, but it is necessary. And the Netanyahu episode serves as a concrete reminder of why this kind of literacy matters — not as an academic abstraction, but as a practical skill for navigating the world we live in today.
At the end of the day, what this episode makes clear is that the battle for digital trust is going to be a long one, and it is going to require effort from multiple sides at the same time. From platforms, which need more effective tools and more transparent policies. From fact-checking institutions, which need greater reach and speed. From AI developers, who bear responsibility for how their technologies are used and abused. And also from every user who shares a video before taking a second to think about where it came from and whether what they are seeing actually makes sense. That one-second pause, multiplied by millions of people, can make a real difference. 🧠
