Real or AI-generated? How to spot fake viral videos before you believe or share them on social media

AI-generated videos are becoming increasingly realistic, making it harder for people to tell genuine footage from artificial content. The problem becomes particularly serious when such videos appear to show disasters, accidents or other major events.

Videos claiming to show the deadly floods in Nepal have been widely shared online. However, not every clip circulating on social media is necessarily genuine. One dramatic video claiming to show the floods was identified as AI-generated, highlighting how easily artificial footage can be mistaken for real events.

Read more - In 2021, when scientists were exploring Antarctica’s Filchner Ice Shelf seafloor, they discovered 'underwater city' with 60 million fish nests & its own natural energy source


As AI video technology continues to improve, viewers need to examine viral footage carefully before believing or sharing it.

Here are some simple ways to check whether a suspicious video could be AI-generated.

Check the video length

The length of a video can sometimes offer an early clue about whether it may have been generated using AI.

The Nepal flood video in question was around 10 seconds long. Some AI video-generation tools, particularly free versions, have traditionally placed limits on the length of videos they can produce.

However, a short video does not prove that it is AI-generated. Genuine videos can also be only a few seconds long, so video length should be treated only as one clue.

Use AI video detection tools

If a video looks unusually polished, strange or too good to be true, viewers can use available AI detection tools to examine it.

Some technology companies have developed tools designed to identify content generated by their AI models. These services can provide useful clues about the origin of a video.

However, AI detection tools are not always accurate. Their results should therefore be considered alongside other evidence rather than treated as conclusive proof.

Slow down the video and check individual frames

One of the simplest ways to examine suspicious footage is to slow it down and look closely at what happens from one moment to the next.

In the Nepal flood video, several unusual changes can be seen. Red containers visible in one section of the footage appear to disappear, while a log seems to appear suddenly without an obvious explanation.

A building and a tree also appear abruptly around the seven-second mark as the camera pans back.

Sudden changes in objects, backgrounds or structures can be warning signs when examining potentially AI-generated footage.

Look for objects that change or disappear

AI-generated videos can sometimes struggle to maintain consistency throughout a scene.

Objects may change shape, disappear and reappear, or develop details that were not present earlier in the video. People, buildings, vehicles, trees and other background elements can also look different from one frame to another.

When checking a viral video, focus on objects in the background as well as the main subject. Small inconsistencies can sometimes be easier to spot when the footage is slowed down.

Ask whether the video makes sense

Another important step is to consider whether everything happening in the footage would be physically possible.

AI-generated videos can have difficulty accurately depicting complex scenes involving fast-moving water, vehicles, people and multiple objects. Look for unnatural movements, distorted details or interactions that do not appear physically plausible.

The more complicated the scene, the more carefully viewers should examine it.

Check where the video came from

The source of a viral video can provide important clues about its authenticity.

Before sharing a dramatic clip, check who posted it and whether the original source can be identified. Look for reports from credible news organisations, official agencies or people who were actually present at the event.

A video being widely shared does not mean that it is authentic.

Compare viral footage with genuine reports

Another useful way to verify a suspicious video is to compare it with confirmed footage and reports from the same event.

In the case of the Nepal floods, genuine images and videos showing people affected by the disaster were already available. Comparing a viral clip with verified coverage can help reveal inconsistencies in locations, weather conditions, buildings or other details.

Checking AI-generated videos

AI-generated videos can spread rapidly on social media because dramatic footage attracts attention and engagement. Some content may be created simply for entertainment, while other videos can contribute to misinformation.

The risk is particularly high during disasters, when people are looking for immediate information and may share emotional or shocking footage without checking its source.

When artificial videos are circulated alongside authentic images and recordings, distinguishing fact from fiction becomes even more difficult.

What to do before sharing a viral video

A few simple checks can help viewers avoid spreading misleading content. Check the video's source, examine suspicious details frame by frame, use AI detection tools where available and compare the footage with credible reports.

No single test can always determine whether a video is real or AI-generated. But combining several checks can make it much easier to identify suspicious footage before sharing it.