Netanyahu Posted a Coffee Shop Video to Prove He's Alive. Grok Called It '100% Deepfake.' The World Can't Agree on What's Real.

Israeli Prime Minister Benjamin Netanyahu posted a video of himself ordering coffee on March 15 to prove he's alive. Within hours, Elon Musk's AI chatbot Grok called it "100% deepfake." Western fact-checkers said the death rumors were false. And millions of people across the Middle East and South Asia weren't sure who to believe.

The story scored a PGI of 7.25 — one of the highest perception gaps of the day. Not because the facts are complicated, but because the same piece of evidence produces opposite conclusions depending on where you're standing.

The Video That Proved Nothing

The sequence started days earlier. After US-Israeli strikes on Tehran killed Iran's former supreme leader Ayatollah Ali Khamenei, rumors spread on Iranian state media that Netanyahu had been assassinated in retaliation.

Then a video appeared on Netanyahu's official X account showing him delivering a war update. Social media users spotted what looked like six fingers on his right hand. The hashtag spread fast: AI-generated. Deepfake. He's dead and they're covering it.

Snopes investigated the original footage from the Israeli Government Press Office. Their verdict: five fingers on each hand. What users saw was the hypothenar eminence — the fleshy bump at the base of the little finger. An optical illusion, not artificial intelligence.

Turkey's Anadolu news agency also called the six-finger claims a hoax.

But that didn't settle it.

Coffeegate

Netanyahu's team tried humor. On March 15, his official account posted a video captioned "They say I'm what? Watch this..." showing the prime minister in a Jerusalem coffee shop, chatting with an aide, ordering coffee, and holding up all 10 fingers to the camera.

"I am dead... for coffee," he said in Hebrew.

The internet dissected every frame. The coffee level didn't drop when he appeared to take a sip. His sleeve seemed to move on its own. A man in the background behaved oddly.

Then Grok weighed in. Responding to a user's query about the video, the AI chatbot wrote: "It's AI-generated. This is a deepfake of Benjamin Netanyahu casually in a coffee shop, talking about Iran/Lebanon ops and protected areas while sipping coffee — nothing like this real event exists."

Asked to confirm, Grok doubled down: "Yes, 100% sure — it's an advanced AI deepfake."

The video was posted on Netanyahu's verified official accounts. Grok — built by Musk's xAI, running on Musk's X platform — was calling the Israeli Prime Minister's own content fake.

Where the World Splits

Here's where the Perception Gap Index picks up the story.

Western outlets — Reuters, Metro UK, Snopes — treated the death rumors as Iranian information warfare. Their framing: Iran's state media spread a false claim, conspiracy theorists amplified it, and AI detection tools (including Grok) gave unreliable results. The story was about misinformation. Middle Eastern coverage framed the IRGC's statement — "If this child-killing criminal is still alive, we will continue to pursue and kill him with full force" — as a legitimate military threat, not mere propaganda. Iran's Revolutionary Guards announced it as part of the "52nd wave of Operation True Promise 4." The deepfake speculation wasn't treated as conspiracy theory. It was treated as a reasonable question during wartime. South Asian media — Hindustan Times, NDTV, Times Now, India Today — covered both sides but gave heavy play to the Grok analysis. Headlines led with the AI's verdict. The framing positioned the story as a genuine epistemic crisis: here's what Grok said, here's what Snopes said, who do you trust?

The PGI dimensions tell the story. Factual divergence scored 7.0 — the highest dimension — because the basic question of whether Netanyahu is alive produces genuinely different answers depending on your source. Actor portrayal hit 8.0: Netanyahu as disinformation victim (West) versus potential casualty of justified retaliation (Middle East). Cui bono reached 7.5 — every narrative transparently serves the interests of whoever's telling it.

The US-Middle East regional pair scored 9.0. That's nearly the maximum possible divergence between two regions covering the same event.

The AI Referee Problem

This story matters beyond the Netanyahu rumors because it surfaces something new: AI tools are now issuing contradictory verdicts on reality, and those verdicts carry weight.

Grok called the video fake. Snopes called the six-finger claim false. Both are positioned as truth-verification tools. Both reached opposite conclusions. And millions of people chose the verdict that matched their existing beliefs.

This is what information warfare looks like in 2026. The question isn't whether a video is real. The question is whether the tools we use to check reality are themselves reliable — and whether we'll accept their answers when they contradict what we already think.

The IRGC didn't need to prove Netanyahu was dead. They just needed the question to exist. And an AI chatbot, built by one of the most prominent figures in American tech, helped keep it alive.

What Happens Next

Netanyahu's coffee-shop video was supposed to end the conversation. It started a new one. "Coffeegate" is now trending across multiple languages, with users analyzing frame by frame whether the video is genuine.

The deeper pattern: in every conflict zone, proof-of-life is becoming a contested category. If AI tools can't agree on whether a world leader's official video is real, the verification infrastructure the internet relies on is breaking down.

Nobody benefits from that except the people who prefer confusion to clarity.

---

This story was scored by the Albis Perception Gap Index — measuring how differently the world frames the same events. See today's most divided stories →