We’ve all seen them: memes or images on-line that are OBVIOUSLY fake. The human with 12 fingers. Your favorite celebrity with three extra chins, wearing a t-shirt with a message you KNOW they don’t support, or beside a sidekick they wouldn’t be caught dead with. What’s to blame? Artificial intelligence (and the geniuses/trolls who wield it!).

 

AI has already proven itself to be incredibly useful in countless fields, but we’re seeing it has the power to amplify mis- and disinformation, spread conspiracy theories, and destroy reputations both virtual and real-life. So, how can you spot AI-generated content? The telltale signs include:

 

1. Wonky text.

Check for misspelled, unclear, or partial text.

 

2. Missing or extra digits and limbs.

Check out hands in the image and make sure all fingers are accounted for and that they (and the arms they’re attached to) rest naturally.

 

3. Blended/helmet hair.

Poor AI struggles with the texture and detail of human hair, so look closely to see if strands are visible or if it all blurs together.

 

4. Asymmetry in architecture.

If a picture of a building looks off—more like an Escher drawing than a photo—AI probably had a hard time getting the geometry right.

 

5. Perfection.

If a photo looks overly filtered (think perfect skin, extra big and bright eyes, or painting-quality landscapes), the robots definitely created it.

 

6. Distorted or unfinished background.

AI is sometimes like a kid who worked really hard on the cover of his final report but then ran out of time actually writing the content. While the subject of an AI image may be spot-on, dig deeper into the background to see half-formed faces, blurry signage, and, frankly disturbing interpretations of real life objects and people.

 

The best advice when faced with a potentially AI-generated image is to take a breath. Before you forward it, post it, or become emotional over its contents, take a long look at it and note any of the signs above. If you’re still not sure—but before you send out a mass mailing—run the photo past Google Lens or TinEye (if it only appears once on the internet, it’s likely fake) or Lenso ai (which offers reverse image searches with facial recognition) or check with Snopes or Reuters Fact Check to see if that image is making the rounds.

 

Worst case? You could just send someone a nice postcard (real image guaranteed!) and skip all the auto-generated drama!