How to Tell If a Photo Is Real: 7 Ways to Spot a Fake

To tell if a photo is real, combine several checks: run a reverse image search for older copies, read the metadata, scan the picture for signs of editing, and ask whether the source and the story add up. If you suspect AI, add a detector as an extra hint. Each method only gives you clues. Real proof exists only when a photo's origin was recorded at the moment it was taken.

This matters well beyond viral news pictures. A phone listed far below market price on Facebook Marketplace, a dating profile that looks a little too perfect, a dramatic image everyone is sharing: editing apps and AI generators now produce fakes that fool trained eyes.

At TrustCamera we build a camera app that gives photos a digital fingerprint at the moment they are taken. That work has taught us exactly where checking a finished image after the fact runs out of road. Each method below links to a detailed guide.

How to tell if a photo is real: the 7 methods at a glance

Seven methods help you tell if a photo is real: reverse image search, metadata, close inspection, AI tell-tale signs, AI detectors, context and proof of origin. You can use the first six on any photo you receive, but they only give you clues. Only proof of origin, created at capture, shows that a file hasn't changed since.

The order matters. Those first steps cost almost nothing and already filter out many fakes. For bigger decisions, such as sending money to a seller you have never met, combine more of them.

Method 1: Run a reverse image search

A reverse image search is the fastest way to tell if a photo is real or simply copied. You upload the picture or paste its link, and the search engine shows where the same or a very similar image has already been published. If a seller's supposedly private photo turns up in an online store or in a news story from five years ago, you have your answer.

Here is how to do it on Google:

1. Go to google.com and click the camera icon in the search box.
2. Upload the image, drag it into the box or paste the image address.
3. Look through the results and pay attention to the oldest appearance.

In Chrome it is even quicker: right-click an image and choose "Search with Google Lens". The Google Search Help page on image search walks through every option. Bing and TinEye offer similar tools and sometimes surface different matches. Our step-by-step guide on how to reverse image search covers all three, on desktop and on your phone.

The limitation: a reverse image search only finds what is already online. A freshly taken photo or a newly generated AI image returns no matches. No match does not mean the image is genuine. It only means the picture was not obviously copied from somewhere else.

Method 2: Check the photo metadata

Digital cameras and smartphones store extra information in every image file, known as EXIF data. It usually includes the camera model, the date and exposure settings, and GPS coordinates if location access was on. If the date contradicts the story, or a supposed phone snapshot came from a professional camera, take a second look.

You can read this data without extra software. On Windows, right-click the file, choose "Properties" and open the "Details" tab. On a Mac, open the image in Preview and show the inspector. For iPhone and Android, see our guide on how to see photo metadata on any device.

Photo metadata is only useful when it is there at all. Many messaging apps and social networks strip it on upload, and anyone who wants to fake it can edit it in seconds with free tools. Missing metadata is therefore no red flag, and present metadata is no proof. When you try to tell if a photo is real, treat it as one clue among several.

Method 3: Inspect the image for signs of editing

Many manipulations give themselves away when you zoom in and scan the image section by section. Give it a few minutes on a large screen. Pay particular attention to these areas:

• Shadows and light: Do all shadows fall in the same direction, and does the brightness of an object match its surroundings?
• Edges: Blurry, jagged or unusually smooth outlines suggest that something was cut out or pasted in.
• Reflections: Windows, glasses and puddles should reflect what is in front of them.
• Repetition: Identical patterns in clouds, leaves or paving often come from cloning parts of the image.
• Lines: Bent door frames or tile joints next to a person are a classic sign of body retouching.

With listings, check the details that have to match the story. Does the charger plug fit the country the seller claims to be in? Is the screen of the "used" phone actually switched on, or does the picture look suspiciously like a manufacturer's product shot? Our guide to Facebook Marketplace scams covers the red flags behind such fake listings.

The limitation: carefully made fakes show none of these traces. Close inspection mainly protects you from quick, sloppy edits, and there are surprisingly many of those in everyday life. For a full walkthrough, read our guide on how to tell if a photo is photoshopped.

Method 4: Look for the typical flaws of AI images

AI generators now create photos that are hard to tell apart from real ones on a phone screen. Many models still have weak spots, though. If you suspect AI, look closely at these points:

• Hands and teeth: too many, missing or merged fingers, teeth that look unnaturally uniform
• Text: signs, labels and license plates with garbled or invented letters
• Background: people and objects that blend into each other or make no sense
• Surfaces: skin without pores and a smooth, almost glossy look

A second approach is labeling. In the EU, the AI Act requires providers of AI systems to mark generated images in a machine-readable way (Article 50). Some generators embed invisible watermarks or provenance data for this purpose.

Do not rely on it. After a screenshot, heavy compression, or with images from open-source models, such labels are often missing. The absence of an AI label alone will not tell you if a photo is real. More signs, labels and tools are in our guide on how to tell if an image is AI generated.

Method 5: Use AI detectors only as a second opinion

Online detectors promise to tell you with one click whether an image is AI-generated. What they actually give you is a probability, not certainty. The result depends on which images the tool was trained on, and compression, cropping or a screenshot can change it considerably.

Germany's Federal Office for Information Security (BSI) points to an example of how hard automatic detection is: in Facebook's 2020 deepfake detection challenge, even the best method reached an accuracy of only about 65 percent. The tools have improved since then, but the core problem remains. A detector mostly recognizes what it has seen before.

Use a detector as an additional hint at most, never as the final verdict. A "likely AI" result is a reason to look more closely. A "likely real" result does not clear the image. With video and voices it gets even harder, as our guide on how to spot a deepfake explains.

Method 6: Check the context and the source

Often it is not the image that gives a fake away, but the story around it. Before you reach for a tool to tell if a photo is real, ask yourself: who published the photo, how old is the account, and does the picture fit everything else you know about the person or the offer? A profile created a few days ago that only shows dream deals deserves suspicion.

For private sales, one simple test helps. Ask for an extra photo with a detail that could not have existed before, for example the item next to a note with your name and today's date. Someone who really owns the item can send that within minutes, and dodging the request is an answer too. Such a photo is no longer proof, though, because current AI generators can render legible text. A short video call in which the seller shows the item live, or seeing it in person, tells you more.

For news images, fact-checkers such as Snopes and the teams at Reuters and AP regularly verify viral pictures and publish where a photo really comes from. Searching their fact-check pages is often faster than any tool.

Method 7: Prove the origin at capture instead of guessing later

All the methods so far share one problem: they try to judge a finished file after the fact. That is why the BSI also names cryptographic methods that take effect at the moment of recording as a countermeasure, since they protect the file from unnoticed changes afterwards.

This is where proof of origin comes in. The open C2PA standard, backed by Adobe, Microsoft and Intel among others, attaches so-called Content Credentials to a file: cryptographically signed information about when and with what an image was created and edited. You can check whether an image carries such information with the Content Credentials verify tool.

TrustCamera pursues the same goal with an approach that needs no special camera. You take the photo or video directly in the app, and uploads from the gallery are not possible. An AI check is meant to catch cases such as someone photographing a screen. The app then stores a digital fingerprint, the capture time and the approximate location, and the capture receives a check code.

Anyone who receives the check code, a buyer for example, enters it at app.trustcamera.com and sees the stored original with its capture time and approximate location. If the image in the listing differs from it, it was altered. It's best to check the code as soon as you receive it. The app for Android and iOS is coming soon. You can already check TrustCamera captures for free in your browser.

It is important to be clear about what such proof shows. It shows that a file has not been changed since it was captured. Whether the scene itself was staged is something no technology can judge. And it only works if the photo was taken that way from the start. For someone else's photo without proof, methods 1 to 6 are how you tell if a photo is real after the fact.

Found a fake photo? What to do next

If you suspect that an image is fake, do not make decisions under time pressure. Scammers push for fast payment because every question increases their risk. These steps help:

1. Don't pay while doubts remain, and above all never send money in advance.
2. Save the evidence: take screenshots of the listing, the profile and the chat before they disappear.
3. Report it: use the platform's report function so the listing or profile gets reviewed.
4. Report fraud if you lost money. In the US you can do this at ReportFraud.ftc.gov.

The US Federal Trade Commission explains the most common warning signs in its guide on how to avoid a scam. Most of its advice applies to online marketplaces worldwide. Lost money on a Marketplace deal? Our guide on what to do if you got scammed on Facebook Marketplace covers refunds and reports.

Frequently asked questions about checking if a photo is real

Can I check if a photo is real online for free?

Yes. You can tell if a photo is real with free tools: reverse image search on Google, Bing or TinEye, the metadata view built into your operating system, and many AI detectors. They only give you clues, though. A photo with a TrustCamera check code can be checked for free in your browser, without an account.

Can you tell if an image was made with AI?

Often, but not with certainty. Typical signs are faulty hands, garbled text and illogical backgrounds, but modern generators make these mistakes less and less. Also check whether the image carries Content Credentials or an AI label, and treat detectors as a second opinion only.

Does EXIF data prove that a photo is real?

No. EXIF data shows what information is stored in the file, such as the camera model and the capture date. Simple tools can edit or delete this information. It is a useful hint when it matches or contradicts the story, but it is not proof of anything.

How reliable are AI image detectors?

They deliver probabilities, not certainty. How well a detector works depends heavily on its training data. With compressed images, screenshots or pictures from new models they are wrong more often, in both directions. Never rely on a detector result alone when money or trust is at stake.

Why does a reverse image search sometimes find nothing?

It only finds images that have already been published somewhere online. New photos and freshly generated AI images have no earlier copies to find. No match simply means the image was not obviously copied. On its own, that says nothing about whether the photo is real.

What does a TrustCamera check code prove?

The code shows that the photo or video was taken with TrustCamera and has not been changed since. You also see the capture time and the approximate location. Whether the photographed scene was staged is something the code cannot judge, and neither can any other technology.

Conclusion: telling if a photo is real means collecting clues

A single test rarely settles it. Combine the quick methods, meaning reverse image search, metadata and close inspection, and take time to check the context before important decisions. The more independent clues point in the same direction, the more confident your judgment can be. If doubts remain, don't pay.

After the fact, authenticity can never be fully proven. It becomes solid only when a photo's origin was recorded at the moment of capture. That is why we built TrustCamera: so that for your own photos you don't have to rely on guesswork, but can simply share a check code.