AI Scam Claim? Check the Source, the Image, and What the Evidence Actually Shows
About this check
If you have just seen an AI scam claim, do not start by asking whether it feels believable. Start by separating the claim from the thing carrying it: What exactly is being asserted, and is the image or screenshot itself authentic?
A finished check should give you more than a reaction. It should identify the claim, show the evidence behind the answer, and make clear when the evidence is not enough for a firm conclusion.
The manual method
First, write the claim in a single, precise sentence. Remove captions, opinions, and predictions that are not part of the claim you want to test. “This image proves X” is not the same claim as “X happened.” Check each one separately.
Next, find the original wording and context. Search distinctive phrases from the claim and compare the results. Look for the date, location, named people or organisations, and the event being described. Open the underlying sources rather than relying only on a repost, headline, or screenshot.
Then compare the evidence. Does a source actually support the claim, or does it merely repeat it? Do independent sources describe the same event? Is the material current, or is an older report being presented as new? Keep a record of the sources that support the claim, contradict it, or leave it unresolved.
For an image, inspect the image separately from the caption. Check whether the visible details fit the stated place and event. Look for an original version or earlier appearances of the image. For a screenshot, read the text exactly as shown and search distinctive wording. Cropping, missing context, and reposting can make an image difficult to assess from the post alone.
This method stops where the evidence stops. A search result is not proof merely because it appears first. A repeated assertion is not independent confirmation. An image may not establish the event described by its caption, and a screenshot may omit information needed to interpret it. If the available evidence cannot settle the question, the responsible answer is inconclusive—not a forced yes or no.
What VOM checks
VOM can accept a claim through voice input and verify claims using real-time web search. Its answer includes a verdict, a confidence score, and the sources used. The verdict can be true, false, or inconclusive, so an unsettled claim is reported as inconclusive rather than forced into a binary answer.
For an uploaded image, VOM assesses whether it is likely authentic, likely AI-generated, or unverifiable. It can also read text inside a screenshot, allowing the text in a forwarded image to be checked without retyping it. These are separate checks: the authenticity assessment concerns the image, while claim verification concerns what the available evidence supports.
VOM independently reviews an AI-generated image or video before approving it for publication. Its article-sharing pipeline still publishes if AI image or video generation is blocked, fails, or is exhausted. This keeps the publication pipeline from depending on a successful generation step.
The useful distinction is simple: manual checking can help you investigate, but it can also leave you with evidence that does not settle the question. A completed verification should show both the conclusion and the basis for it—and should say when no firm conclusion is justified.