Fact-check: AI-generated video falsely shows young Pakistanis urging people to avoid protests
The footage also appears to feature recurring faces, with two men presented as different individuals appearing remarkably similar
Updated Monday Sep 28 2026
A video circulating online ahead of the Pakistan Tehreek-e-Insaf's (PTI) planned long march towards Islamabad purportedly shows young Pakistanis urging people to avoid protests involving violence and vandalism.
The video is AI-generated.
Claim
On September 25, a user on X (formerly Twitter) shared a 34-second video featuring young women and men delivering what appears to be a public service message against participating in protests that lead to vandalism and road blockades.
The clip ends with the message: “Do not take part in violence, vandalism, road blockades or clashes with law enforcement.”
The video was shared ahead of PTI's planned long march, which was initially scheduled for September 27 but later postponed to October 4.
The footage was presented as a genuine recording of young Pakistanis delivering the message.
Fact
The video was in fact generated using artificial intelligence (AI), according to a media forensics expert who examined the footage and results from two AI-detection tools.
Dr J. Rosenbaum, a media forensics expert at the Royal Melbourne Institute of Technology (RMIT) in Australia, reviewed the clip at Geo Fact Check's request and identified several visual inconsistencies commonly associated with AI-generated videos.
“The biggest indication that this is AI is the woman in purple, the patterning on her clothes has artifacts that are consistent with AI,” Rosenbaum told Geo Fact Check via messages.
Rosenbaum also noted that the appearance of a watch worn by the last man in the video shifts in a way that seems inconsistent with reality.
Dr Rosenbaum also pointed to another man's fingers, which appear to merge together in some frames. “I can now confidently say that this is AI-generated,” the expert explained.
The footage also appears to feature recurring faces, with two men presented as different individuals appearing remarkably similar.
Geo Fact Check separately analysed the clip using Hive Moderation, an AI-content detection platform. The tool assigned the footage an aggregated score of 99.9%, indicating a high likelihood of AI generation.
The team also examined the footage using DeepFake-o-Meter, a detection tool developed by the University at Buffalo. Its detection models indicated a “high” probability that the video was AI-generated.
Verdict: The video is AI-generated and does not show genuine footage of young people delivering the message.
Follow us on @GeoFactCheck on X (Twitter) and @geo_factcheck on Instagram. If our readers detect any errors, we encourage them to contact us at [email protected]