See if you can spot an AI deepfake with our test
Summary
Researchers from the University of Aberdeen and Australian National University have studied whether people can be trained to identify AI-generated fake human faces, known as deepfakes. They found that, by learning to notice subtle features in faces, people can improve their ability to tell real images from AI fakes within a short training period.Key Facts
- AI can now create very realistic fake human faces called deepfakes.
- Early AI mistakes, like extra fingers, helped detection but are now rare.
- Researchers trained people to look for six clues: symmetry, proportionality, attractiveness, distinctiveness, expressiveness, and memorability.
- AI often produces faces that are too perfect, more attractive, more generic, less emotional, and less memorable.
- AI struggles more with creating realistic non-white, older, or younger faces.
- Training involved showing participants mixed real and AI faces and explaining which were which.
- Participants improved from about 40% correct guesses to significantly higher accuracy after training.
- The study used images generated by StyleGAN3, a leading AI face generator.
Read the Full Article
This is a fact-based summary from The Actual News. Click below to read the complete story directly from the original source.
