4 minute read
How to spot AI images and videos
You often cannot tell an AI image by looking anymore. Instead, check where it came from: who posted it first, whether trusted news outlets show the same thing, and whether it carries Content Credentials, a label that tools like ChatGPT, Adobe and Google attach to generated files. Treat shocking images that appear only on social media as unconfirmed.
Check the source, not the pixels
- Who posted it first? Use Google Lens or another reverse image search to find the earliest copy.
- Do trusted outlets show it? Real major events are photographed by many people from many angles.
- Is there a label? Drop the image into Content Credentials Verify. A label proves it was made or edited with AI. No label proves nothing; screenshots remove it.
- Does it make you angry or scared? Fakes are made to spread fast. Wait a day before sharing.
Clues that still sometimes help
- Text in the background that is garbled.
- Hands, jewelry, glasses and ears that do not quite match.
- Skin that looks airbrushed, lighting that comes from two directions.
- In video: mouths that do not match the words, blinking that looks off.
These clues get rarer with every new AI model, so the source check matters more.