Skip to content
Check it yourself

Verify Assistant: Check a Claim, Image, or Screenshot With Sources

Every verdict carries a confidence score and the sources behind it.

About this check

If you searched for “verify assistant,” you probably want a direct answer to something you just saw: Is this claim true? Is this image authentic? What does this screenshot actually say?

A finished verification should give you more than a yes or no. It should show a verdict, indicate how confident that verdict is, and cite the sources used to reach it. When the evidence does not settle the claim, the responsible answer is inconclusive—not a forced answer in either direction.

The manual way to verify a claim

Start by separating the exact claim from the surrounding post. Write down what is being asserted, who or what it concerns, and any date, location, number, or quotation attached to it. A vague question produces vague results; a precise sentence is easier to check.

Next, search the key phrase in quotation marks, then search the important names, figures, or events without the quotation marks. Look for reporting close to the original event, official statements, public records, research, or other sources that can directly support or contradict the claim.

Check the date and context of each result. An old article can be accurate while still being irrelevant to a new post. A real quotation can also be reused to suggest something its original context did not mean. Compare what multiple sources actually say rather than counting how many posts repeat the same wording.

For a screenshot, read the text carefully and search distinctive phrases from it. Check whether the account, publication, date, and surrounding context can be found elsewhere. If the screenshot contains a claim, verify that claim separately; the existence of an image does not establish the truth of what it says.

For an image, inspect the original context and look for an earlier publication, a source, or a version with additional details. A copied image can be genuine but miscaptioned, while an altered or generated image may look plausible without proving the event it appears to show.

Finally, write down what the evidence supports, what it contradicts, and what remains unknown. Stop short of certainty when the available sources are weak, conflicting, or insufficient. Manual checking can become difficult when a claim is new, a source has disappeared, an image has been separated from its context, or the available evidence does not answer the specific question.

What VOM returns

VOM accepts voice input for submitting a claim to verify. It uses real-time web search and cites the sources behind each verdict. Its response includes a verdict, a confidence score, and the sources it used. A verdict can be true, false, or inconclusive, so a claim that the evidence cannot settle is reported as inconclusive rather than forced into yes or no.

You can also use VOM to assess an uploaded image as likely authentic, likely AI-generated, or unverifiable. It reads the text inside a screenshot, allowing a forwarded image to be checked without retyping it.

Verification is part of a broader assistant experience. VOM holds natural back-and-forth conversations for everyday questions, learning, brainstorming, writing, and planning. It can turn a written idea into a generated image, and it offers local, world, and trending feeds of circulating claims. VOM answers in the language being used across its eight supported languages.

The useful distinction is simple: manual verification gives you a process for investigating a claim yourself; VOM presents a verdict, confidence score, and cited sources for the checked claim, while leaving room for an inconclusive result when the evidence does not settle it.

Keep reading