Claim: A video shows an Iranian soldier targeting and launching a missile at an American helicopter and bringing it down.
Fact: The footage is AI-generated.
On 11 June 2026, X user Tariq Masood Butt posted a video allegedly showing the Iranian army shooting down a US helicopter by a missile,with the caption:
غیر متوازن جنگ کی ایک زبردست مثال! اطلاعات کے مطابق ایران صرف 150 ڈالر کے کم لاگت ہتھیار استعمال کر کے اربوں ڈالر مالیت کے امریکی فوجی طیاروں کو نشانہ بنا رہا ہے۔ یہ واضح یاد دہانی ہے کہ سستی ٹیکنالوجی آج مہنگی فوجی برتری کو کیسے چیلنج کر رہی ہے۔
“An astonishing example of asymmetric warfare! According to reports, Iran is targeting billion-dollar American military aircraft using low-cost weapons that cost just $150. This is a stark reminder of how inexpensive technology is challenging costly military superiority today.”
Screenshot from X post
This claim was shared in the context of reports about a US helicopter that was downed on 9 June 2026, at the Strait of Hormuz, allegedly by Iran. Iran’s deputy foreign minister told Al Jazeera that they did not deliberately down the American chopper.
This article is not investigating the incident in early June, it only fact-checks the video.
Fact or Fiction?
This claim has been widely fact-checked by international organisations also accredited by the International Fact-Checking NetworK (IFCN). Several organisations, including Lead Stories, France 24 and AFP Fact-Check found that the video is AI-generated. Soch Fact Check’s independent analysis found the same.
AFP Fact Check debunked this video through a mix of technical forensics and visual anomalies. DeepFake-O-Meter results for the scanners AVSRDD, DSP-FWA and AltFreezing all show a high probability of AI-generation in the video content. Visually the inconsistencies detected was the American flag on the soldier’s arm in the first frame of the video contradicting it was an Iranian soldier.
Further, reports show that the helicopter was shot down by an Iranian drone, AFP pointed out that in the video it was a foot soldier in the video that contradicts reports of the US defence system. Hence, proving the video was AI-generated.
France 24 also debunks the video through visual flaws, observing an American flag patch on the alleged Iranian soldier, and the wrong weaponry being used. France 24 also noted that the silhouette of the helicopter does not match the distinctive one of an Apache helicopter.
American flag found in X clip
Soch Fact Check also reverse-searched key frames of the video, to find an official source of the video. But found no official footage that was released by the Iranian government or the US government.
In the clip, one can hear and see the helicopter crash before the missile hits the chopper. This desynchronisation in audio can indicate some form of manipulation and videos are compressed when posted on Meta platforms causing blurriness and out of sync audio.
Soch Fact Check compared the footage to an official BBC footage of a Russian helicopter being hit by an Iranian missile. The first point of the observation is that the fire seems unnatural in the footage being fact-checked whereas, as seen below in a real life collision, smoke should typically be a mixture of black, grey and white than just purely black that stayed throughout till the body of the helicopter crashed into the water.
Another evidence we observed in the BBC video was on collision with the fire till it hit the ground whereas in the X clip the fire at collision is observed as a spark at 00:11 seconds and is gone instantly. The claim is in comparison not showing any realistic elements of what will happen if a missile hits a helicopter causing it to be believed it can be AI-generated.
Comparison between the US helicopter being hit by an Iranian missile (left) and the BBC’s footage of a Russian helicopter being downed by a Ukrainian missile.
Results from AI Detection Tools
Furthermore, for supporting evidence we used Hive Moderation as the next AI detection scanner, which involves multiple tools like visual, audio and text moderation. It checks if an image or text was created by AI.
We ran the image through Hive Moderation’s AI-Generated Content Detection tool, used specifically for photos.The tool analyzes visual patterns, textures, lighting, and facial features amongst other characteristics and provides a probable source of the image and videos.
The aggregate video result is 49.5% for the likelihood of AI-generation. Sometimes, in low resolution videos it is a normal occurrence to have a low average score when it comes to an overall average score of AI-generation video.
However, within the breakdown of these results, specific frames at 00:06 to 00:09 second, showing a heightened likelihood of AI-generation at 91.7%. This is the highest score indicated by the scanner for artificial intelligence. The 91.7% starts at the moment of the launch, and to the missile’s smoke trail, flagging it as AI-generated.
Results generated by Hive Moderation
To further investigate if the video was altered, we ran it through DeepFake-o-Meter, an AI-based tool that detects manipulated or synthetic media, particularly deepfakes. It uses multiple detection models to analyse visual and audio cues that may indicate tampering. The results were as follows:
Results generated by DeepFake-o-Meter
One of the detectors, AVSRDD, displayed the highest probability with 100%. Its functionality looks at both sound and lip movements in a video to check if the speech is real or fake. It proves that video includes artificially overlaid or synthetically manipulated audio.
The next detector was AltFreezing giving a percentage of 86.5%.This framework specializes in identifying spatial and temporal anomalies within facial structures. The high score presents a high probability that the facial structure of the soldiers’ have been altered.
Thirdly, the CFM detector concluded 47.7%. It flags anomalies in facial texture and blending. The score is near to the baseline of 50% score, suggesting at surface level these features look convincing, and AI-generation might not have been picked up by this scanner.
Furthermore, WAV2LIP showed us a result with 36.8%.This detector explicitly evaluates whether mouth shapes organically map to spoken phonemes. This indicates that the scanner did not pick up any AI-generation specifically when it came to the lip movement corresponding with the audio that is being heard. Even though this score is low, and might contradict the analysis, it can be because of the low resolution of the video. Hence, we cannot discredit the other high scores that have been given by other scanners.
In conclusion, with the results from two online detectors and our individual analysis of the visual inconsistencies in the video, Soch Fact Check confirms that AI was used to create the video.
Virality
The video circulated on Facebook with the most popular post gaining 7.9 likes and 868 reshares.
The post was shared on X in two instances, with one post gaining 1,298 views and another gaining 144K views.
Conclusion: The video is AI-generated.
To appeal against our fact-check, please send an email to appeals@sochfactcheck.com