How To Tell If An Image Is AI Generated

Telling AI images from real ones gets harder every year, but four kinds of checks, used together, still give a good answer in most cases. Start with the quickest.

1. File metadata and Content Credentials

Many AI tools write their name or a tag into the image file, and some add signed Content Credentials. Real photos often carry camera details instead. In Chrome, right-click an image and use AI Image Detector (made by Patrick Bushe, who runs this site) to read this information; it gives a verdict such as "AI Generated" or "Likely Real Photo," with the fields behind it. Upload files with Content Credentials to Content Credentials Verify for the full history.

If there's no metadata, which is common for anything shared on social media, move on; that result doesn't point either way.

2. Platform labels

Instagram and Facebook show "AI info," LinkedIn shows a "CR" icon for Content Credentials, TikTok and YouTube label disclosed or detected AI content, and Pinterest labels some Pins "AI modified." Labels help when present but are often missing.

3. The image's history

Run a reverse image search with Google Lens, TinEye or Bing. Look for the earliest copy and who posted it. A real event photo usually appears from several sources and angles; an AI image often traces back to a single account or an AI art community.

4. Close inspection

• Hands, teeth, ears and jewelry.
• Text on signs, packaging and clothes.
• Reflections, shadows and light direction.
• Background objects and people.
• Repeated patterns or melted shapes.

Then decide

One strong signal, such as AI metadata or a confirmed original from an AI account, is usually enough. Several weak signals together are suggestive. If nothing settles it, call the image unverified rather than real or fake.

See what still works in 2026 and free tools to check an image.