Are people good at spotting AI-generated faces?

No, research indicates people are overconfident in their ability to spot AI-generated faces, performing only slightly better than chance.

What makes AI-generated faces hard to detect?

Advanced AI faces are hard to detect because they appear unusually average, highly symmetrical, and statistically typical, lacking obvious visual glitches.

What are the risks of not being able to spot AI faces?

The inability to reliably spot AI-generated faces increases vulnerability to scams, fake profiles, and fabricated identities.

Home / Technology / AI Faces Fool Human Eyes: Can You Tell Real From Fake?

AI Faces Fool Human Eyes: Can You Tell Real From Fake?

19 Feb

•

Summary

Most people are overconfident in identifying AI-generated faces.
Advanced AI faces appear unusually average and symmetrical.
Misplaced confidence in spotting fakes increases vulnerability to scams.

AI Faces Fool Human Eyes: Can You Tell Real From Fake?

A recent study from the University of New South Wales in Australia reveals that people are consistently overconfident in their ability to identify artificial intelligence (AI)-generated faces. What were once obvious visual glitches in fake images have been replaced by an unnerving perfection, with AI faces often appearing unusually average, highly symmetrical, and statistically typical.

Researchers found that both average individuals and 'super-recognisers' performed only slightly better than chance when distinguishing between real and AI-created faces. This misplaced confidence, however, remains high, according to co-author Dr. James Dunn. This makes individuals and organizations more susceptible to scams, fake profiles, and fabricated identities, as the assumption that one can 'just tell' when a profile picture is fake is no longer reliable.