AI RESEARCH
Human-AI Ensembles Improve Deepfake Detection in Low-to-Medium Quality Videos
arXiv CS.AI
•
ArXi:2603.14658v1 Announce Type: cross 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 professionally captured DF40.