A video appears in your WhatsApp group, Instagram feed or YouTube recommendations. The caption says it shows a shocking event happening right now. Friends are sharing it, comments are piling up, and the clip looks convincing. Should you believe it?
Pause before you share. A viral video can be genuine but old, real but filmed somewhere else, edited to change its meaning, or created with artificial intelligence (AI). The most reliable way to check it is not to rely on one clue—it is to look for the original source, test the claim and compare independent evidence. A 2026 guide from the UK House of Commons Library makes the same essential point: visual content must be checked for both authenticity and context.
1. Find the original source
Ask three questions: Who first posted this? When was it posted? What did the original uploader say it showed? A repost, anonymous account or forwarded message may have removed the details needed to understand the clip.
Search a distinctive phrase from the caption, the name of the place, or a short description of what happens in the video. Look for an earlier upload and compare its caption, date and context. If a clip said to show an event today was already online years ago, that is a strong warning that the current claim may be misleading.
2. Take a screenshot and reverse-search it
There is no single universal reverse-video search that proves a clip is genuine. But you can often search still frames from it to find earlier appearances of the same scene.
- Pause the video at a clear moment—ideally one with recognisable buildings, signs, people or landmarks.
- Take a screenshot, then open Google Lens or Google Images and search using the image.
- Look for older versions, original uploads, news reports or fact-checks that show the same frame.
- Repeat with another frame if necessary. A single matching image may not tell the whole story.
Google explains that its “About this image” feature, where available, can provide context such as when similar images may first have been seen and which other pages use them. Search results are clues, not a final verdict: an image may be cropped, edited or absent from search indexes.
3. Check the date, place and details
A video can show a real event and still be shared with a false caption. Check whether the location and date match what the post claims.
- Landmarks: Do buildings, road signs, shop names, language or landscape match the claimed location?
- Weather and time: Do shadows, daylight and visible weather fit the claimed date and place?
- Sequence: Do cuts, audio and actions make sense, or does the clip omit what happened immediately before or after?
- Independent evidence: Are there other videos, photographs or credible reports of the same event from different sources?
Do not treat one odd-looking detail as proof of fakery. Videos can be compressed, edited for ordinary reasons or recorded in difficult conditions. Look for evidence that can be independently checked.
4. Search for fact-checks and reliable reporting
Search the central claim—not just the viral caption—alongside terms such as “fact check”, “old video” or the name of the location. You can also use Google Fact Check Explorer to look for claims that independent fact-checkers have already investigated.
For major events, check whether credible news organisations or relevant official agencies have independently confirmed the basic facts. Several websites repeating the same unverified social-media post do not count as several independent confirmations. Trace reports back to their original evidence wherever possible.
Google’s guide to spotting misinformation also recommends checking image context, comparing news coverage, consulting fact-checkers and using maps to assess location claims.
5. Could the video have been made or changed with AI?
AI tools can generate realistic scenes, imitate voices or alter existing footage. Possible warning signs include lip movements that do not match speech, inconsistent reflections, distorted background details, strange text or abrupt changes in sound. But these clues are not conclusive: real videos can also contain visual glitches, and AI-generated media is improving quickly.
Use provenance and corroboration rather than relying on appearance alone. Check whether the original creator or platform has labelled the clip as AI-generated, look for credible reporting, and investigate any available content credentials or watermarks. In October 2026, Google announced broader access to its SynthID Detector, which can help identify watermarks in media made with Google AI and participating partners. A result from a tool like this has a limited scope: it does not establish that every video without a detected watermark is real, nor does it verify the truth of the caption.
Likewise, do not upload a clip to an unfamiliar “AI detector” and treat its score as proof. Tools can make mistakes. The source, timeline, context and independent evidence still matter.
6. Use this 60-second check before sharing
- Source: Can I identify the original uploader?
- Date: Is this footage actually from the date being claimed?
- Place: Do the visible details match the location?
- Search: Does a reverse-image search of a frame reveal an older or different context?
- Corroboration: Is there independent evidence, not just copies of the same post?
- Claim: Does the footage really show what the caption says it shows?
If you cannot answer these questions, the responsible choice is simple: do not forward it as fact. You can ask for the original source or say that the video has not yet been verified. If it could put someone at risk, avoid spreading it while you check.
Quick questions
Does a reverse-image search prove a video is fake?
No. It may reveal that a frame appeared earlier or in a different context. That is an important clue, but the full clip and its claim still need to be checked.
Can I spot every AI-generated video by looking closely?
No. Visual errors can be clues, but there is no dependable visual shortcut for every case. Check origin, context, corroborating evidence and any trustworthy provenance information.
What if no search results appear?
No results do not prove that a video is original or genuine. It may be new, private, cropped, edited or simply not indexed. Continue checking the source and the claim, or wait for reliable confirmation.
The takeaway
Virality measures how far content spreads—not whether it is true. Find the original source, search frames from the video, check its date and location, look for independent confirmation and be cautious with AI-detection claims. When the evidence is incomplete, withholding a share is better than helping misinformation travel.
Sources and further reading: House of Commons Library: How to verify images and videos (2026); Google India: Spot misinformation online with these tips; Google Search Help: Learn more about an image; Google DeepMind: SynthID Detector update (October 2026).
