See if you can spot an AI deepfake with our test

Distinguishing between real photographs and AI-generated deepfakes is becoming a significant challenge, as artificial intelligence technology rapidly advances. Previously, spotting computer-generated images, often used by fraudsters, was easier due to obvious AI errors like an extra finger or other peculiar flaws.

However, AI systems are continuously learning and improving, rendering these tell-tale signs less reliable. "Training on visual artifacts, like looking for a sixth finger or odd earrings, has had limited success, partly because the AI is getting too good, and fraudsters may avoid using pictures with obvious flaws anyway," explained Professor Amy Dawel.

Professor Dawel, director of the Australian National University Emotions and Faces Lab, is leading an international research team. Their work focuses on determining if people can be trained to detect AI imposters. The current answer is yes, but it requires a more sophisticated approach than simply looking for glaring mistakes. The ability to spot AI fakes now depends on a subtler understanding and analysis of the images.

“

Training on visual artifacts, like looking for a sixth finger or odd earrings, has had limited success, partly because the AI is getting too good, and fraudsters may avoid using pictures with obvious flaws anyway.