AI-generated deepfake faces have become so realistic it is increasingly difficult for people to tell them apart from real humans.
Researchers at the Australian National University Emotions and Faces Lab have begun training people to spot AI-generated faces, with their work published in the scientific journal PNAS on Tuesday and replicated by colleagues abroad.
Lead researcher Amy Dawel said telling real faces apart from AI-generated ones had become more difficult as technology rapidly evolved.
"AI faces are advancing rapidly," Associate Professor Dawel told AAP.
"This is an incredibly challenging area ... it's like fighting computer viruses."
The training offered by the ANU team directs people's attention to global qualities that differ between AI and human faces.
"AI faces tend to be more symmetrical, proportional and attractive, but without training, we often think these are markers of being human," Assoc Prof Dawel said.
"Fake faces are hyper-average. They're more proportional ... they're too perfect."
The researchers trained people to spot AI-generated faces by drawing their attention to six perceptual qualities: distinctiveness, memorability, proportionality, symmetry, attractiveness and expressiveness.
"When you look at any individual AI-generated face, it can be difficult to tell, but when you see a group of AI faces, they start to look alike," Assoc Prof Dawel said.
The ability of all participants in the study to spot AI faces dramatically improved, researchers found.
"We've shown our training is effective for some of the most convincing fakes available, StyleGAN faces," Prof Dawel said.
"Now we need to find out whether that training generalises to other AI-generated faces."
The team hopes its research can help people navigate increasingly complex digital environments, including fraud and staying safe online.
"AI is here and it's here to stay. There's no escaping it," Assoc Prof Dawel said.
Researchers at the University of Victoria in Canada have replicated the ANU study.
The ANU Emotions and Faces Lab is encouraging people interested in undertaking AI face detection training to register for future studies.
13°C