They say seeing is believing, but thanks to artificial intelligence (AI) we can no longer assume the pictures and photos we see online are true.
Today, anyone can use free AI chatbots and other online tools to make photos that look like the real thing. Of course, some are super obvious and easy to identify. For example, you might see a silly image like this as a meme in your social media feed.
The tricky ones though, are those that look completely legitimate. This picture of a woman walking with two rainbows behind her, for example, was made by a computer.
In this case, the harm in believing it’s real is harmless—but what if you booked a vacation based on these photos:
If you did, one of two things would happen:
- Your credit card information would be stolen.
- Your vacation location would absolutely NOT look like what is pictured.
Why? Because this resort doesn’t exist. All of these images were made in about 30 seconds using an AI image generator. Whoever posted these either wanted to trick you into spending money on a vacation spot that wasn’t nearly as nice as the pictures, or they flat out wanted to trick you into giving them your payment information.
From social media posts claiming something is true, to fake ads, there are billions of these visual lies out there in the world—with more being created every day. You can make sure you don’t fall for them by training your eye and leveraging technology to help.
How to spot an AI image with your own eyes
If you look closely, computer-generated pictures often (but not always) have specific “tells” that don’t quite match reality. Watch for:
- Incorrect human details. AI often struggles to get arms and legs correct. Look at the hands and fingers, strange angles, or even extra limbs of all the people pictured.
- Too much perfection. In real life, things aren’t tidy. Hairs end up out of place, fabric crinkles, windows are dirty, and small variations happen everywhere. In AI’s world, images can come out TOO smooth, too symmetrical—too perfect.
- Texture repeats. AI can struggle with complex patterns and textures, leading to noticeable odd repetition.
- Weird lighting or shadows. Do the lighting angles and shadows line up like they should? Is the light too bright where it shouldn’t be?
- Bad text. Letters and numbers in AI images can often end up like gobbledygook or just misspelled or mixed up.
Check out these sources for more tips on spotting AI images: Britannica Education, BBC Bitesize.
Use an online detector
If you open a browser and search for “free AI image detector”, you’ll spot quite a few websites allowing you to upload a file and let technology (maybe even AI?) do the checking for you. For example, when we searched, we spotted:
We’re tossing these out here just as a few examples, but know AI detectors are out there and could be a helpful tool.
Test yourself!
Research says AI-generated images are so real looking, people fall for them half the time—and with AI continuing to improve, it will only get more difficult to spot the fakes. How do you think you measure up? Test yourself by taking these free online quizzes:
