How to tell if a photo is AI-generated: 8 simple checks anyone can do
AI images are getting more realistic every month. These quick checks, from looking at details to using free tools, help you spot fakes before you share them.
A shocking photo appears in your feed. It looks real, but is it? AI image generators can now create pictures that fool most people at first glance. The good news is that you don’t need special skills to check. Here are eight simple steps that work for most images.
1. Zoom in on hands and fingers
Hands are complicated, and AI has long struggled with them. Look for extra or missing fingers, fingers that merge together, or strange joints. Newer tools are much better at hands, so a normal-looking hand doesn’t prove a photo is real, but a strange one is a strong warning sign.
2. Read any text in the image
Signs, labels, T-shirts and number plates often give AI away. Look for letters that turn into nonsense, misspelled words or text that doesn’t match the setting.
3. Check the background
AI usually focuses on the main subject and gets lazy with the rest. Look for faces in a crowd that melt into each other, buildings with impossible shapes, or objects that blend into the scenery.
4. Look at light, shadows and reflections
- Do shadows fall in the same direction?
- Do reflections in mirrors, glasses or windows match the scene?
- Is the skin unusually smooth and glossy, like a plastic doll?
5. Do a reverse image search
Upload the picture to Google Lens or TinEye, or right-click it in Chrome and choose “Search with Google Lens”. You may find the original source, an older version, or fact-checkers who have already looked at it. If a “breaking news” photo has no trace anywhere else, be careful.
6. Check for Content Credentials
Many companies, including Adobe, Microsoft, Google and OpenAI, support a standard called Content Credentials (also known as C2PA). It attaches a record of how an image was made. On some websites and apps you’ll see a small “CR” icon you can click. You can also upload an image to the free Content Credentials Verify website. If credentials exist, they may say the image was AI-generated. If they don’t exist, that tells you nothing either way, because credentials are easily removed.
7. Use “About this image”
Google Search’s “About this image” feature can show when an image first appeared online and, for images made with Google’s AI tools, whether it carries an AI watermark (SynthID).
8. Think about the story
Ask simple questions. Who posted it first? Is it from a trusted news organisation? Does it seem designed to make you angry or scared? Fake images spread fastest when they trigger strong emotions.
The best fake detector is a pause. Before you share, take ten seconds to check.
What about AI detector websites?
Many websites claim to detect AI images, but their results are often wrong in both directions: they can label real photos as fake and miss real fakes. Use them as just one clue, never as final proof.
Quick checklist
- Zoom in on hands, text and backgrounds
- Check light, shadows and reflections
- Reverse image search
- Look for Content Credentials
- Consider who posted it and why
If you still can’t tell, the safest choice is simple: don’t share it.
