Why a picture feels like evidence
Most screenshots people send are honest, which is part of the problem. We learn that a picture of a payment usually means a payment, and the habit hardens into a rule.
Psychology explains why the rule feels so solid. In 1999 Rolf Reber and Norbert Schwarz showed that statements which were easier to read were judged true more often. A screenshot is about as easy to read as information gets: a layout you see every day, a green tick, a familiar font.
Pictures also lend weight to claims they cannot support. In experiments by Eryn Newman and colleagues in 2012, adding a photo that proved nothing made people more likely to judge a claim true. The picture did not have to be evidence. It only had to be there.
What a screenshot shows, and what it doesn't
What it shows
- What someone chose to put in front of you
- A layout you recognise from an app you use
- That an image of this exists
What it doesn't
- That money reached your account
- That a chat happened, or happened in full
- Whose account, ID or balance it really is
- That a badge, rating or review is current or earned
Notice who made the picture. A screenshot is evidence prepared by one side of the deal, at a time of their choosing, with every chance to crop or change it. No court would accept that as proof, yet we accept it in a chat in seconds.
Who it fails, and how
Sellers meet it first. Delhi Police's cyber cell describes fraudsters on OLX sending fake payment screenshots made with spoofing apps, and using fake identity documents of national institutions to gain confidence. The FTC warns US sellers about buyers who send a fake payment notification and hope you post the item before you notice.
Buyers meet it as social proof: screenshots of glowing reviews, of follower counts, of a verified badge. In August 2024 the FTC finalised a rule banning fake reviews, including AI-generated ones, paying for reviews, and buying fake followers. A market has to be large before a regulator writes a rule for it.
Making a convincing image now costs almost nothing
Editing a number on a screenshot never took much skill. What has changed is the rest of the picture. In a study by Sophie Nightingale and Hany Farid, people sorted real faces from AI-made ones with 48% accuracy, worse than a coin toss, and rated the made-up faces as slightly more trustworthy.
So looking harder is not a plan. Sometimes a fake has a wrong font or an odd time stamp, but a screenshot with no visible flaw proves nothing, and the better the fake, the more confident it makes you.
What real proof looks like
We think the useful test is one question: did I fetch this, or was I handed it? Proof you are handed was prepared for you. Proof you fetch comes from a source the other person does not control.
- For a payment, open your own bank or payment app and look at the balance and history.
- For a profile, badge or review, open the platform yourself and look it up there, not in a picture of it.
- For a person, use a check you start, live, that could not have been prepared in advance.
- For a claim like verified or licensed, follow it to a page run by whoever did the verifying.
Questions
Can't I spot a fake screenshot if I look closely?
Won't an honest person be offended if I don't trust their screenshot?
Is a screen recording any better than a screenshot?
We built Realr so proof is a page you open yourself rather than a picture someone sends: you look up a person's verified profile at realr.me, on our site, not in a screenshot of it.
Look someone upSources
- Reber and Schwarz (1999): Effects of perceptual fluency on judgments of truth, Consciousness and Cognition (PubMed)
- Newman et al. (2012): Nonprobative photographs (or words) inflate truthiness, Psychonomic Bulletin and Review (PubMed)
- Delhi Police Cyber Cell: OLX fraud, modus operandi and safety tips
- FTC: Selling stuff online? Here's how to avoid a scam (2022)
- FTC: Final rule banning fake reviews and testimonials (Aug 2024)
- Lancaster University: AI-generated faces rated more trustworthy than real faces (Nightingale and Farid, PNAS 2022)



