Human-AI Ensembles Improve Deepfake Detection in Low-to-Medium Quality Videos
This work addresses the problem of improving deepfake detection for real-world applications, particularly in non-professionally produced videos, by showing that human-AI collaboration is necessary, though it is incremental in combining existing methods.
The study tackled deepfake detection in realistic video conditions by comparing human and AI performance, finding that humans outperformed AI, especially in low-to-medium quality videos where AI accuracy dropped to 0.537 while humans maintained 0.784 accuracy.
Deepfake detection is widely framed as a machine learning problem, yet how humans and AI detectors compare under realistic conditions remains poorly understood. We evaluate 200 human participants and 95 state-of-the-art AI detectors across two datasets: DF40, a standard benchmark, and CharadesDF, a novel dataset of videos of everyday activities. CharadesDF was recorded using mobile phones leading to low/moderate quality videos compared to the more professionally captured DF40. Humans outperform AI detectors on both datasets, with the gap widening in the case of CharadesDF where AI accuracy collapses to near chance (0.537) while humans maintain robust performance (0.784). Human and AI errors are complementary: humans miss high-quality deepfakes while AI detectors flag authentic videos as fake, and hybrid human-AI ensembles reduce high-confidence errors. These findings suggest that effective real-world deepfake detection, especially in non-professionally produced videos, requires human-AI collaboration rather than AI algorithms alone.