Video of passenger being offloaded from plane is AI-generated

Claim: A clip shows a Pakistani woman protesting her removal from an airplane and demanding to know why she was offloaded.

Furthermore, the clip claims the woman has done nothing wrong, did not come late and still got offloaded from the planee. 

Fact: The clip is AI-generated. 

Recently, a clip was shared on Facebook with the claim that the Pakistani “regime” was offloading passengers. This video is apparently from a Pakistani airport immigration counter and features a woman arguing and protesting against being offloaded from her flight. 

The Urdu clip states her saying:

“Why did I get offloaded? Is it because I am poor? This is in justice, why are you all stopping me? I have done nothing wrong.

Fact or Fiction?

Soch Fact Check initially suspected the video is fake, mainly as the woman’s mouth movements do not match what she is saying, and her voice is out of sync. 

We then noted several indicators of the video being AI-generated. It is zoomed in and filmed only from the front; whereas an actual incident would likely have been shot from multiple angles. The woman’s cheeks and forehead appear unnaturally smooth and airbrushed. 

Soch Fact Check reached out to Shaur Azher, a lecturer who teaches sound design and sound recording at the University of Karachi and the Shaheed Zulfikar Ali Bhutto Institute of Science and Technology (SZABIST). He also works as an audio engineer at our sister organisation, Soch Videos, and specialises in mixing and mastering audio.

An analysis was carried out on the audio of the video being fact-checked.Two sound samples were compared, both of which consist of arguments between civil aviation and passengers: Sample A is the audio from the video in the claim and sample B is from real footage of Allama Iqbal International Airport, Lahore. The report stated below revealed that the audio of Sample A is fake  : 

Observations & Findings

1. Frequency Spectrum Behavior

Sample A (Doctored Audio):

1. Maximum variable frequency energy peaks at approximately 5,835 Hz, after which

significant spectral cluttering is observed.

1. High-frequency energy rapidly decays despite the subject allegedly screaming.

III. Frequency energy levels remain unnaturally low, inconsistent with a loud vocal event

recorded at close-to-mid distance.

1. Such behavior is indicative of synthetic filtering, bandwidth limitation, or AI voice

generation, rather than natural microphone capture.

Sample B (Real Footage):

1. Variable frequencies extend across the full audible spectrum, with balanced energy

distribution.

1. Minor spectral cluttering is present but remains consistent with real-world environmental

noise.

III. Loud vocal events show expected high-frequency excitation, consistent with natural

human screaming.2. Clipping & Gain Behavior

Sample A:

1. No audio clipping observed, even during peak screaming moments.
2. Smartphone video recordings typically lack hard limiters; therefore, screaming at the

recorded distance should have resulted in peaks exceeding 0 dBFS and noticeable

clipping.

III. Dialogue and screaming remain at similar amplitude levels, suggesting artificial gain

control or post-processing.

1. This behavior strongly contradicts real mobile device recording characteristics.

Sample B:

1. Minor clipping observed during loud vocal events.
2. This aligns with expected behavior from consumer smartphone microphones when

exposed to sudden high SPL (sound pressure levels).

III. Gain response appears natural and uncontrolled, supporting authenticity.

3. Dynamic Range & Compression Analysis

Sample A:

1. Clear evidence of post RMS compression.
2. Dynamic range appears flattened, with reduced contrast between quiet and loud

passages.

III. This indicates post-production processing or AI rendering, not raw capture.

Sample B:

1. No signs of post-production compression.
2. The dynamic range is uneven and unbalanced, typical of raw, unprocessed field

Recordings.4. Environmental Acoustics (Reverb & Spatial Characteristics)

Sample A:

1. Hall reverb is present but appears artificial and exaggerated.
2. The reverberation profile resembles a large, empty hall, inconsistent with a populated

international airport environment.

III. In real airports, crowd density significantly reduces reverberation due to sound

absorption by human bodies and objects.

1. The acoustic response suggests synthetic spatial modeling rather than real-world

reflection patterns.

Sample B:

1. Minor reverb is detectable during peak screaming moments.
2. Most dialogue exhibits little to no reverberation, consistent with a crowded terminal

environment.

III. Acoustic behavior aligns with expected sound diffusion and absorption in a populated

airport.

Conclusion

Sample A is highly likely to be AI-generated and/or post-processed audio.

Sample B is consistent with genuine real-world audio recorded at Jinnah International Airport.

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:

Soch Fact Check then compared a genuine video/image found via Google search of a passenger from Jinnah International Airport to a keyframe from the video in the claim. A side-by-side comparison of the two images, highlights several inconsistencies in the video in question, weakening the claim that the clip was recorded at a Pakistani airport. Further, the background of the video being fact-checked does not look like any major airport in Pakistan, most of which the Soch Fact Check team have viewed.

1. Uniform Color and Design

• In the genuine reference image from Pakistan, an immigration officer is wearing a light grey uniform, which is their official uniform, clearly identifiable by its cut, color, and insignia placement.
• In contrast, the officer standing behind the woman in the video is wearing a dark navy-blue uniform, which does not match the known uniform standards of Pakistani immigration or airport security staff.
• Pakistani immigration staff do not operate in dark blue uniforms, particularly not in passenger-facing immigration or security areas.

2. Insignia and Accessories

• Pakistani airport personnel typically display clear badges, shoulder insignia, and standardized nameplates consistent with FIA or ASF regulations.
• The officer in the video lacks these recognisable markers, further suggesting he is not part of Pakistan’s airport authorities.

3. Airport Architecture and Layout

• Pakistani airports generally feature more compact immigration halls, distinct signage styles, and locally recognisable design elements.
• The terminal seen in the video appears larger, more modern, and architecturally inconsistent with Pakistani airport interiors, especially in terms of ceiling design, lighting layout, and spatial openness.

The visual evidence strongly suggests that the officer and the airport shown in the clip are not Pakistani. This significantly undermines the claim that the video was filmed at a Pakistani airport and supports the conclusion that the clip has been AI- generated. 

Virality

The video was shared with the claim on X here. 

It was shared on Facebook with the same claim here. 

It was also shared on Tiktok. 

This claim was found on instagram here as well. 

Conclusion:  

An investigation carried out by Soch Fact Check, based on publicly available information and open-source techniques to verify visuals circulating on social media, concludes that this viral clip is AI- generated.