Want to spot a deepfake? Look for the stars in their eyes

In an era where deepfake technology is prevalent, a new study presented at the Royal Astronomical Society's National Astronomy Meeting suggests a novel way to detect AI-generated fake images. By analyzing the reflections in a person's eyes, researchers found that inconsistencies in the reflections between the left and right eyeballs can indicate a deepfake. This method draws parallels to how astronomers study galaxies by quantifying the reflections in the eyes. The research, led by University of Hull MSc student Adejumoke Owolabi and Professor Kevin Pimbblet, uses techniques like the CAS and Gini indices to compare the reflections and identify discrepancies. While not foolproof, this approach provides a valuable tool in the ongoing battle to identify deepfakes, offering a unique insight into distinguishing real images from AI-generated fakes based on the reflections in the eyes.