WordSig: QR streams enabling platform-independent self-identification that's impossible to deepfake

View on arXiv ← Back to list

Authors: Andrew Critch

Published: 2022-07-15 17:23:01+00:00

AI Summary

WordSig is a protocol enabling video participants to cryptographically sign their spoken words using QR codes, verifiable by viewers independently of content distributors. This creates a trusted connection between speaker and viewer, resistant to deepfakes.

Abstract

Deepfakes can degrade the fabric of society by limiting our ability to trust video content from leaders, authorities, and even friends. Cryptographically secure digital signatures may be used by video streaming platforms to endorse content, but these signatures are applied by the content distributor rather than the participants in the video. We introduce WordSig, a simple protocol allowing video participants to digitally sign the words they speak using a stream of QR codes, and allowing viewers to verify the consistency of signatures across videos. This allows establishing a trusted connection between the viewer and the participant that is not mediated by the content distributor. Given the widespread adoption of QR codes for distributing hyperlinks and vaccination records, and the increasing prevalence of celebrity deepfakes, 2022 or later may be a good time for public figures to begin using and promoting QR-based self-authentication tools.

Key findings

The paper proposes a novel approach to combat deepfakes using QR codes and digital signatures. The system is designed to be platform-independent, relying on standard cryptographic techniques for verification. While not a complete solution, it offers a proactive, user-empowering method for authenticating spoken content within videos.

Approach

WordSig overlays a stream of QR codes onto videos. Each QR code contains a timestamped digital signature of the preceding audio segment, verifiable using the speaker's public key. This binds audio to visual identity, making deepfakes difficult.

Datasets

UNKNOWN

Model(s)

ECDSA (Elliptic Curve Digital Signature Algorithm)

Author countries

USA

← Previous