AI ethics laws and regulations, artificial intelligence legal standards and policy | Photo Credit: anyaberkut

Amid the West Asia conflict, a clip of Israeli Prime Minister Benjamin Netanyahu addressing the nation began circulating on X, and users quickly noticed something wrong: his hand appeared to have six fingers. Iranian state media seized on the moment, amplifying claims that Netanyahu was dead and that the government was using AI to fake his appearances. To prove he was alive, Netanyahu released a video of himself at a Jerusalem café. But Grok, the AI assistant embedded in X, flagged that one as potentially generated by AI. A third video followed. Then a fourth. None was conclusively verified.

Misinformation on social media is not simply propaganda, but a tornado of synthetic and real content cleverly stitched together and backed up by AI slop that even professional fact-checkers cannot reliably verify.

A FLOOD OF SYNTHETIC CONTENT

Swasti Chatterjee, Deputy editor (news) BOOM, an independent digital journalism initiative that fights misinformation and explains issues, told businessline that she has spent eight years tracking misinformation across wars, elections, and disasters. She says the West Asia conflict is more overwhelming than anything she has ever dealt with.

One example described by Chatterjee: a viral image of an Iranian missile bearing the text ‘In memory of Epstein Island victims’. When Boom ran it through both free and paid AI detection tools, the results were inconclusive. The team fell back on reverse image search—the only reliable method available—to trace the image’s origin and verify whether it was AI-generated.

Rakesh Dubbudu, CEO of Factly, a civic tech, fact-checking, and public data research firm, told businessline that he estimates AI-generated content now accounts for roughly one-fifth of all conflict-related misinformation. He identifies three distinct categories emerging at scale: Fully synthetic imagery of explosions and airstrikes; audio deepfakes layered over real footage of Indian officials; and — most alarmingly — real videos being branded as AI-generated.

“The danger of actual real videos being branded as AI — for the first time, we are seeing it at scale during this conflict.” — Rakesh Dubbudu, CEO, Factly.

WHEN AI ASKS AI

Perhaps the sharpest irony of this conflict’s information war is that people are turning to it for help. In previous crises, users would tag fact-checking organisations when they encountered suspicious content. That behaviour has changed.

“People were not tagging fact-checkers. They were tagging Grok,” Chatterjee says Grok, was repeatedly giving wrong answers — forcing Boom to monitor the chatbot’s responses in real time and battle it with publishing corrections alongside their regular fact-checks.

“Our job became double-fold. We were not only producing fact checks, but also replying to Grok’s replies — saying, Grok just got it wrong. An AI chatbot cannot replace fact-checkers.”— Swasti Chatterjee, Deputy News Editor, Boom Bangla

Dubbudu echoes the warning. “Don’t go asking Grok, because it can make mistakes,” he says. “Keep a tab on major fact-checking agencies. They follow some methodology — they show evidence. At least you get additional information from there.”

WIDER THAN THE WAR

Misinformation about conflicts extends well beyond the fringes of the internet. Two examples from the West Asia conflict illustrate just how widely fabricated narratives can travel before being caught. A video circulated claiming that an Indian crew in the Strait of Hormuz had agreed to pay in Chinese yuan to be allowed to pass — a claim the Indian Ministry of External Affairs was forced to publicly debunk. Boom considered fact-checking it independently, but doing so would have required a response from the Chinese side, which proved extremely difficult to obtain, and so they chose not to proceed.

The USS Abraham Lincoln became another flashpoint. Pro-Iranian handles circulated claims that it had been bombed and was being withdrawn because it was damaged. The carrier was repositioned around the same time, US Central Command took to social media to deny that claim — this illuminates how misinformation exploits every genuine ambiguity in an active conflict, making it harder to dismiss and easier to believe.

WHAT YOU CAN DO

Rakesh Dubbudu and Swasti Chatterjee offer practical advice. First: pause before sharing information online. The viral life cycle of misinformation runs on speed — and that speed is exactly what its creators are counting on.

Second, cross-reference social media posts with news links from established outlets such as The Hindu and The New York Times. If something significant had happened — a world leader killed, a warship sunk — they would have covered it. If a claim seems too dramatic to believe, that instinct is worth trusting. Seek it out in a credible publication before forwarding it to anyone else.

Third, and most importantly: Do not outsource fact-checking to an AI chatbot. As this conflict has shown, even the most widely used AI tools are producing false positives on ambiguous content where accuracy matters most. Human fact-checkers — with their institutional accountability have been at the forefront when misinformation runs rampant.

“Just depending on detection tools is not enough,” Chatterjee says. “We have to run a double eye. We have to figure out the visual discrepancies ourselves on.” In a conflict where a six-fingered prime minister can reach millions before anyone thinks to question it, the human eye is the last reliable line of defence.

Published on April 2, 2026