Protecting world leaders against deep fakes using facial, gestural, and vocal mannerisms

Abstract

Since their emergence a few years ago, artificial intelligence (AI)-synthesized media—so-called deep fakes—have dramatically increased in quality, sophistication, and ease of generation. Deep fakes have been weaponized for use in nonconsensual pornography, large-scale fraud, and disinformation campaigns. Of particular concern is how deep fakes will be weaponized against world leaders during election cycles or times of armed conflict. We describe an identity-based approach for protecting world leaders from deep-fake imposters. Trained on several hours of authentic video, this approach captures distinct facial, gestural, and vocal mannerisms that we show can distinguish a world leader from an impersonator or deep-fake imposter.

Human Perception of Art in the Age of Artificial Intelligence

“…On this thinking, DALL•E's capacity to produce works that observers tend to prefer over human artworks can be explained by the fact that its training dataset comprises images of artworks that are broadly considered to be aesthetically pleasing. This is in line with recent research on 'deepfakes', wherein AI-generated faces not only fool observers with their hyper-realistic nature, but are also associated with enhanced perceptions of trustworthiness 2,14,15 . These findings raise critical concerns about the exact nature of the cognitive processes that could be targeted and manipulated using generative-AI, and therefore, about its large-scale deployment without detailed investigation.…”

Qualitative Failures of Image Generation Models and Their Application in Detecting Deepfakes

“…In this study, we focused on still images. However, for videos, additional indicators beyond those outlined here, such as motion and optical flow, as well as the synchronization of lip, face, and head movements over time, can also be significant factors [3]. Educating individuals on the cues outlined in this paper may aid in combatting deepfake proliferation.…”

Identification of new threats to the national security of the state

“…Tested on several hours of authentic video, the identification-based deepfake detection approach was found to capture distinct gestures, facial and voice features. In this way, the use of deepfakes against world leaders during elections or during armed conflicts is minimized (Boháček and Farid, 2022).…”

“…An increased number of cases of cyberattacks, which may have military or political goals, has been noted. Farid (2021), Boháček and Farid (2022) considered issues related to state and individual protection measures against deepfakes, ethical use of deepfakes as a weapon. Particular concern about the use of deepfakes against world leaders during armed conflicts has been noted.…”