How to Verify a Photo Online

Compare file-level evidence with the Image Comparison Tool, then follow the photo and visual verification learning path to test source, date, place, and caption as separate claims.

Photo verification is not one magic lookup. It is a series of independent checks that test the caption, source, time, place, and file history. When the evidence arrives inside an interface capture, use the screenshot verification workflow to separate the account, text, embedded image, and claimed time.

Write the claim before checking it

Turn the post into a testable sentence: “This image shows event X in location Y on date Z.” Then investigate each part separately. A real image can still support a false caption.

The five-check workflow

1. Source: identify the account or publisher and look for an original upload.
2. Reuse: reverse-search the full image and distinctive crops for earlier appearances.
3. Time: compare upload dates, weather, shadows, signage, and known event timelines.
4. Place: inspect landmarks, road markings, language, terrain, and storefronts.
5. File: examine dimensions, compression, metadata, and editing clues without treating metadata as infallible.

Look for caption drift

A photo may move from a dated news report into a later post that assigns it a new place or event. Finding an older, well-documented use can disprove the later caption even if the image itself is authentic.

Keep an evidence table

Claim element	Supporting evidence	Contradicting evidence	Status
Source	Named photographer	Earlier unattributed copy	Open
Date	Publisher timestamp	Same image found two years earlier	Contradicted
Location	Matching landmark	No conflict found	Probable

Walk through a caption-verification case

Imagine a photograph shared with the claim that it shows a current protest in a named city. Begin with the claim, not the visual. The image must support at least the event, place, and time. Reverse search may reveal that the same frame appeared years earlier with a different caption.

Street signs and architecture may test the location. Weather records or shadows may help, but only when their limits are understood. The file itself may add metadata, although a social copy often lacks it.

If an older use is found, verify that it is truly the same frame. Compare fixed details such as hand positions, folds, vehicle placement, signs, and background geometry. A photograph from the same event can look similar without disproving the new claim. Save the older page, its visible date, and an archive capture if available.

Write a verification note with calibrated language

• Claim tested: State the exact event, place, time, or person being checked.
• Confirmed evidence: List only observations supported by the image or reliable sources.
• Contradictions: Record the specific detail that conflicts with the caption.
• Unresolved points: Name what could not be established and why.
• Conclusion: Use supported, contradicted, or unresolved instead of a broad “real” or “fake” label.

This structure prevents an authentic photograph with a false caption from being described as a fabricated image. It also makes the reasoning reviewable because every conclusion points back to a particular claim and piece of evidence.

If the claim could cause harm, have a second reviewer repeat the key comparison from the saved sources. Agreement should come from the evidence trail, not from seeing the first reviewer's conclusion.

Continue with The Limits of What a Reverse Image Match Can Prove. An evidence guide for turning image matches into careful findings without treating a search result as proof of identity or ownership. For another evidence check within the same investigation, use How to Check If an Image Is AI-Generated.

Verification and provenance references

• Google: About this image context and early uses
• C2PA Content Credentials specifications