NBC News99%
Experts sound alarm as new AI-generated videos of Iran war spread across social media 94%
4/17/2026, 11:58:37 PM
BS Summary: This video contains 27 faulty reasoning types, including Negativity Bias, Appeal to Authority, and Appeal to Emotion, with Hasty Generalization as the most egregious example at 24.2% saturation with 144 hits. Analysis detected 1,441 faulty-reasoning hits from 596 analyzed words, generating a BS Score of 91.1% and a BS Rank of 94% (1,006 of 16,813 videos). This video is worse (more manipulative) than 94.00% of the video peer group.
A massive explosion on the Israeli port of Hifa.
A blast outside an Iranian apartment building.
A news report about key US allies.
>> Every piece of what I'm about to walk you through is documented.
>> You'd be forgiven for thinking these are real videos from the Iran war.
But every single one of them, according to forensic analysis, is fake.
And in the case of the erroneous news report,
>> but here is the reality. debunked by the actual real life reporter after her post.
At least some of the fake videos of her were taken down.
Part of a wider information war playing out online made possible by artificial intelligence.
>> The quantity of AI uh photos, images is unprecedented.
>> Masa Alamarani is the associate director of Witness, a human rights organization working on the impact of AI on video evidence.
>> The lie travels much faster than the truth.
So even when we do see things debunked, people still have a vested interest in believing in the lies of that generated content.
>> She points to last month's alleged US air strikes on an Iranian elementary school in Minab, southern Iran, killing more than 170, many of them children.
Already a tragedy with news organizations verifying the very real human toll.
But the regime used AI in an apparent attempt to amplify the horror, posting this image of a bloody backpack tweeted by the Iranian embassy in Austria.
We uploaded the image to Google's Gemini and quickly discover the image you've shared shows signs of being edited or generated by Google AI.
A visual analysis also reveals several common indicators of AI generation.
When asked, Google directed us to a blog post about its AI verification technology, which says the company has quote long invested in ways to provide you with helpful context about information you see online.
Then there are the parody videos.
>> You're my favorite toy.
>> Not meant to deceive, but to evoke emotions.
Propaganda poking fun at a deadly serious situation.
If you block me, then I block you.
>> In this case, the Iranian embassy of South Africa posting an AI generated music video about Trump's new strategy in the street of Hormuz.
There's plenty of AI generated misinformation coming from supporters of the US too. hours
after President Trump announced the successful rescue mission over Easter weekend.
>> And he was extracted from enemy territory.
>> This fake image went viral, even reposted by some US lawmakers purporting to show the rescued US airmen before they took those posts down.
At all means, experts are increasingly sounding the alarm about a concept called the liars dividend.
Just the very fact that we know that there is AI content in the information ecosystem makes people doubt things are real and it works to the benefit of bad actors.
While Alamar Darnney says fake content is being generated by all sides for the moment in the war for attention.
>> I do believe the regime is winning the narrative war in the information space.
>> Why they are winning? because they have been studying the space and they've never had more raw material to work with,
>> especially when that raw material can be generated in seconds and is fake. Aaron McLaclin, NBC News.
>> We thank you for watching and remember, stay updated on breaking news and top stories on the NBC News app or watch live on our YouTube channel.
Analysis
Hover over highlighted words in the article to view the associated bias or fallacy analysis.