Faking the front line: How artificial intelligence is overriding war perceptions

The video was ten seconds long. It showed Tel Aviv's skyline convulsing under missile fire, towers disappearing into smoke, sirens tearing across a city already half-destroyed. The captions said Iranian ballistic missiles had overwhelmed Israel’s air defences. Within hours, it had millions of views. One major Indian newspaper published it as fact. It was entirely fabricated. That was March 2026. But the machinery behind it had been running for over a year. A review of documented cases by several international news agencies and a range of fact-checking organisations across three war zones has identified more than twelve incidents between January 2025 and March 2026 in which AI-generated video was used to reshape public understanding of active conflicts, in Gaza, Ukraine and the Iran-Israel theatre.

The same playbook, three war zones

The method is consistent. A synthetic clip, often posted openly on a TikTok or Instagram account that labels its content as AI-generated, gets stripped of that label, fitted with a breaking-news caption timed to a real event, and released. By the time a fact-checker responds, the damage is done.

Gaza

In Gaza, a dramatic explosion video captioned as an Israeli air strike on civilians was traced by The Real Fact Hunter to a TikTok account called @everythingai97, which had clearly marked the clip as AI-generated. An October 2025 video of a so-called Gaza Martyrs Museum, complete with statues of well-known victims, was tracked by Misbar to an Instagram account that openly identifies as an AI content creator. In December, two separate clips of children and tents submerged in floodwater went viral as humanitarian evidence. AFP Fact Check found one scored 99.9 percent on AI-detection platform Hive Moderation.

Ukraine-Russia

Ukraine saw a different operation. In November 2025, DW Fact Check identified a wave of videos showing tearful young men in Ukrainian uniforms pleading not to be sent to the front, shared by pro-Russian accounts as proof of forced conscription. Several carried visible Sora watermarks, the logo of OpenA’s video tool. Around the same time, dozens of clips claimed to show mass Ukrainian surrenders near Pokrovsk. Ukraine’s Center for Countering Disinformation issued a formal warning, calling it a coordinated information attack. Independent fact-checkers found faceless soldiers, fused fingers and no corroborating battlefield evidence.

Iran war

The Iran-Israel escalation produced the most extreme material. Reuters Fact Check consulted three forensics experts on the March 2026 Tel Aviv footage, all confirmed AI origin, citing warped solar panels, a flag that changed design mid-zoom and audio that lacked real-world reverberation. An Indian outlet, BOOM ran the same clips through two detection tools; both returned approximately 99 percent synthetic probability.

Virality outperforms fact-check

Vineet Khunger, Co-founder of IndieVisual, a leading AI technology-led video production company, says the barrier to creating convincing synthetic news footage has collapsed entirely. “It’s become a bit too easy for a bad actor sitting alone at home to use foundational AI video models to create clips that look like real news,” he told WION. “Someone with nefarious intent can figure out ways to easily fool people on social media. We’ve already got issues from both angles, some people believing footage that’s fake, and on the flip side, other people doubting what they see even if it’s real.” Vineet says trained professionals can spot the tells, faked CCTV footage that cuts every seven to eight seconds, the current generation limit for consumer AI video models, along with glitchy artefacts and unnatural visual clarity. But he is candid about the limits of that expertise at scale.

“Given that we’re scrolling past so much video, it’s not humanly possible to go deep into the authenticity of every clip we see. Social media rewards engagement, a fake clip designed for virality will almost always outperform a post pointing out that it was fake.”

When institutions fell for it

The fakes did not stay on the fringes. Just days ago, a sitting UK Member of Parliament shared an AI-generated image depicting rows of shrouded bodies said to be schoolgirls killed in an airstrike on a school in Minab, Iran. Five separate fact-checking organisations identified it as synthetic. A leading Indian newspaper article briefly treated an AI missile-strike clip as real footage before issuing a correction.

The structural problem is not that detection tools are inadequate. It is that the strip-and-amplify pipeline moves faster than any newsroom can respond. The label proving a video is fake is removed before the video goes viral. The correction arrives after the damage is complete. On the question of who bears responsibility, Vineet is direct. “Everyone, the production industry, the media industry, the government, that should be pressurising the tech companies to include clear safeguards around what can be generated,” he said. “The government should step in with context-specific policies. AI video in the news space should be heavily regulated. AI video on social media should be clearly labelled as such. AI has many positive use cases, but these safeguards can give us a more trustworthy, transparent and safer environment for everyone.”

Fourteen months of documented fakes across three conflict zones have exposed a gap that technology alone cannot close. To put it simply, AI has not changed what information warfare is for. It has changed who can wage it, how fast, and with what degree of plausibility. For the millions who never saw the fact-check, the fake was simply the news.

Disclaimer: WION takes utmost care to accurately and responsibly report ongoing conflicts in West Asia involving Israel, Iran, the US, Gulf nations, and non-state actors like Hezbollah, Hamas, Houthis, Islamic State, and others. Claims and counterclaims, disinformation and misinformation are being made online and offline. Given this context, WION cannot independently verify the authenticity of all statements, social media posts, photos, and videos.