New paper accepted
Mirko Cassu, Luca Guarnera, Pasquale Caponnetto, and Sebastiano Battiato, we asked ourselves: can cognitive biases influence decisions in forensics and digital forensics, especially when using generative or manipulative AI tools, such as deepfakes?
We tried to answer that question and, in addition, we baptized what we believe to be a matter of increasing concern: “impostor bias“, as we called it, is the a priori tendency to doubt the authenticity or validity of multimedia content such as audio, images, and video because of the increasing accuracy of #generativeai products. This bias can lead to errors or false accusations, compromising the reliability and credibility of forensic evidence.
In our article, previewed below on arXiv and currently being processed on Forensic Science International: Digital Investigation, we go into detail about the most common and dangerous cognitive biases in forensics and digital forensics, as well as methods to counter the increasingly common cognitive biases due to deepfakes and massive multimedia production via Generative AI.