Deepfakes are systematically worse at faking emotive facial expressions than they are at replicating neutral, boring faces.
April 25, 2026
Original Paper
Interpretable facial dynamics as behavioral and perceptual traces of deepfakes
arXiv · 2604.21760
The Takeaway
Human perception and machine detection both excel at spotting fakes during moments of high emotion. Deepfake algorithms struggle to maintain realistic facial dynamics when a subject is crying or laughing. This degradation provides a clear signal for identifying manipulated videos. The most believable deepfakes are those where the person shows no emotion at all. Focus on the subtle movements of the eyes and mouth during intense expressions to tell if a video is real.
From the abstract
Deepfake detection research has largely converged on deep learning approaches that, despite strong benchmark performance, offer limited insight into what distinguishes real from manipulated facial behavior. This study presents an interpretable alternative grounded in bio-behavioral features of facial dynamics and evaluates how computational detection strategies relate to human perceptual judgments. We identify core low-dimensional patterns of facial movement, from which temporal features charact