Video of soldiers apologizing for injustices against the people of Thira Valley was generated by AI

Claim: The video shows a Pakistan Army soldier admitting that the people of the Thira Valley are being treated unfairly by the state.

fact: Clips are generated by AI.

On January 25th, an X user posted Soch fact check have consistently fact confirmed in past Posted video (archive) Statement by a Pakistani soldier admitting the injustices committed against the people of the Tirah Valley. The post reads:Our neighbors openly admit it, “We are the oppressors.” we are not humans. May God destroy us. ”

#TilaValley”

The video shows the soldier saying: [translated from Urdu]:

“I am standing here in the Thira Valley right now, and the injustice that is being done to these people is very unjust. I swear to God, this is not a question of justice. But what can we do? May Allah punish them in such a way that those who are complicit in this cannot stand up. That is why those who oppress these people are called oppressors.”

Residents of Tirah Valley, Khyber Pakhtunkhwa Province, It has started evacuate their home in fear imminent military operations against Pakistan Tehreek-e-Taliban. This area has also been affected strict The snowfall added to existing difficulties for residents caught in the process of leaving their homes.

Fact or fiction?

Soch fact check Analysis of the video revealed that it was generated by AI.

The soldier’s hair and mustache are airbrushed, but his skin is unnaturally smooth and free of blemishes and textures, indicating that AI was used. Additionally, the bottom left of the clip says “AI generated,” but users don’t notice it in the caption of the post.

Additionally, a TikTok user named “@804imrankhan154” also appears in the clip.Taking a hint from this, Soch fact check I reverse searched the keyframes from the clip and found that this was actually the one that was uploaded. TikTok User (archive) January 22nd.

This TikTok video is labeled as “Contains AI-generated media” (see image below).

TikTok videos include labels that alert users that they are generated by AI.

Analysis of the above TikTok user account show that he uploaded other Videos labeled as AI-generated:

TikTok users have uploaded several AI-generated videos.

Additionally, a backward search also returned the same Facebook. video (archive), similar to the TikTok clip, this appears to be the earliest version (uploaded on January 22nd), but without any overlaid text or music. This Facebook user also appears to have uploaded multiple AI-generated clip on his account.

Soch fact check We ran both Facebook and TikTok versions through AI detection tools to ensure that the clips circulating online were indeed generated by AI. hive moderation is one such tool. For Facebook clips, the probability is 99.5%, meaning “the video likely contains AI-generated or deepfake content.”

Hive moderation results for the Facebook version of the same clip.

For the same TikTok video, Hive returned a 99.8% probability, indicating that “overall, the video is likely to contain AI-generated or deepfake content.”

Hive moderation results for the Facebook version of the same clip.

moreover, When I run both versions, Zhuque AI Discovery Assistant It has also become clear that AI is being used. Videos from Facebook and TikTok received consistently high scores of 79.03% and 86.44%, respectively, and were labeled as “likely to be generated by AI.”