Fooled Twice – People Cannot Detect Deepfakes But Think They Can

Hyper-realistic manipulation of audio-visual content, i.e., deepfakes, presents a new challenge for establishing veracity online content. Research on the human impact addressing both behaviors in response to and cognitive processing of remains sparse. In pre-registered behavioral experiment ( N = 210), we show that (a) people cannot reliably detect (b) neither raising awareness nor introducing financial incentives improves their detection accuracy. Zeroing in underlying processes, find (c) are biased towards mistaking deepfakes as authentic videos (rather than vice versa) (d) overestimate own abilities. Together, these results suggest adopt ``seeing-is-believing'' heuristic deepfake while being overconfident (low) The combination renders particularly susceptible be influenced by inauthentic