Claim: A video shows a government representative confirming that Facebook will be banned in Pakistan in 2026.
Fact: All the clips are AI-generated as they carry a Veo watermark.
On 15 December, a Facebook post (archive) shared a video apparently showing a government representative confirming that Facebook will be banned in Pakistan in 2026. It features a man standing at a podium, addressing the media. The description of the post states [translated from Urdu into English]: “Facebook is being completely shut down in Pakistan by 2026. The government has officially confirmed this. #Pakistan #viral #news #karachi #islamabad”.
Another similar post (archive) included a video of a man claiming that Facebook will be shut down by 20 December.
Fact or Fiction?
Soch Fact Check looked into the claim as all the viral videos include a Veo watermark which indicates they are AI-generated. An initial keyword search did not yield any conclusive evidence that Facebook will be shut down in Pakistan in December of this year or in 2026.
We did, however, find an older press release from the Pakistan Telecommunication Authority (PTA) from May 2010 that issued directives to ban Facebook following the Lahore High Court (LHC) decision. “The Lahore High Court had ordered the Telecom ministry to ban the social networking website following the Prophet Muhammad (PBUH) caricatures controversy,” The Express Tribune reported in May 2010.
It is important to note that social media platforms, including Facebook, have faced temporary shutdowns or restrictions, particularly during protests or religious processions in Pakistan over the last few years. However, there are no credible reports that the government recently announced that Meta’s platform would be shut down.
As of the writing time, Facebook is accessible, debunking the claim that it was going to be banned in December of this year.
Veo-generated videos
Video 1 (archive): Analysing this clip shows it has all the signs of being AI-generated.
The Urdu text appearing on the news ticker on the screen, as well as the microphones in front of the speaker, is garbled and nonsensical, a telltale sign of AI being used in its creation as it fails to generate coherent text.
Garbled Text on the ticker and microphones.
The reporters’ hands, visible at 0:05, distort and morph (as highlighted below), in some moments appearing to have more than five fingers, another indication of AI’s usage.
Highlights showing distorted hands.
Moreover, the clip has a “Veo” watermark in the bottom-right corner, confirming that it was created with this software. Veo is an AI-generation software by Google. This year, the tech giant announced it would be adding this watermark to videos generated by the software.
A Veo watermark appears at the bottom-right corner of the video.
Soch Fact Check also ran the clip through multiple AI detection tools. Two tools, Hive Moderation and Zhuque AI Detection Assistant, yielded low probabilities of 0.4% and 1.12%, respectively, indicating that the tools suggest the content is not AI-generated.
However, Deepfake-O-Meter, an initiative of the University at Buffalo, indicated a high probability that it was AI-generated. Nine out of twelve algorithms yielded scores above 50% (as shown below), which confirms it is fake.
Results from Deepfake-O-Meter.
Therefore, Soch Fact Check’s analysis, the presence of a Veo watermark, and the results from Deepfake-O-Meter confirm that the video is AI-generated.
Video 2 (archive): This clip, too, has all the signs of being AI-generated. For instance, garbled text appears on the screen ticker and the microphones in this video also.
Towards the very end of the footage, we see one hand with only four fingers, and another one next to it appears blurred and distorted. Both are highlighted in the screenshot below.
Hands distorted and a Veo watermark.
Moreover, a Veo watermark also appears in this clip, similar to the watermark in the first video, confirming that it was created using the software.
Soch Fact Check also ran this clip through multiple AI detection tools. Hive Moderation in this case yielded a high probability of 99.6%, which means “overall, the video is likely to contain AI-generated or deepfake content”.
Results from Hive Moderation for the second video.
Deepfake-O-Meter’s seven algorithms also yielded probabilities above 50%, which indicated that the video is likely AI-generated, as shown below:
Results from Deepfake-O-Meter.
Only one tool, Zhuque AI Detection, however, gave a score of 11.49%, meaning that, according to this tool, this video is not AI-generated.
The results from two out of the three AI detection tools, a Veo watermark, and Soch Fact Check’s analysis confirm that the clip is AI-generated.
Virality
The first clip was viewed at least 81,000 times and received 1,700 reactions.
The second clip garnered 21,00 views.
This user also shared other similar videos that have a Veo watermark here, here, here, here, and here.
Conclusion: Videos claiming to show a government representative announcing that Facebook would be shut down are fake. All the videos carry a Veo watermark, confirming that they are AI-generated.
Background image in cover photo: Dawn
To appeal against our fact-check, please send an email to appeals@sochfactcheck.com