IDEMIA Leads French Government’s Counter-Deepfake Consortium

IDEMIA is assisting French government efforts to counter the growing threat of deepfakes, the company has revealed. Deepfakes that manipulate video and audio recordings to create realistic but false representations have risen to prominence as potential fraud threats, with the rapid advancement of AI technology making it increasingly difficult to distinguish between real and fake media.

In response to the issue, the French Service National de Police Scientifique (SNPS) identified the need for advanced forensic tools to combat deepfakes. This led to the launch of the A Prototype Assessment Toolbox for Forensic Experts (APATE) project in 2022, funded by the French National Research Agency (ANR). The APATE project aims to develop innovative and reliable deepfake detection methods, providing forensic experts with the necessary tools to bring deepfake cases to court.

The APATE project involves five organizations working together to keep deepfake detection methods up-to-date and effective. The research focuses on detecting deepfakes through various means, such as analyzing low-level traces of noise, blur, and compression, and using speaker recognition and anti-spoofing techniques. The goal is to create a comprehensive toolbox that forensic experts can use in legal proceedings to provide objective and accurate assessments of deepfake content.

IDEMIA plays an important role in this collaborative effort by leading the consortium of partners, including SNPS, LRE-EPITA, École Polytechnique, and ENS Paris Saclay. The consortium combines expertise from each organization to develop and refine the deepfake detection tools. IDEMIA says the SNPS is actively involved in all project decisions, testing and implementing the tools to ensure their practical applicability in real-world forensic scenarios.

Since the project’s inception in September 2022, significant progress has been made in cataloging deepfake attacks and understanding the techniques used to create them.

In its post revealing its work in the project, IDEMIA asserted that it is only “in its preliminary stages.” So far work has primarily revolved around cataloguing deepfake attacks. The next phase involves learning about various deepfake detection models using inventoried databases and state-of-the-art techniques. This will enable the creation of the first deepfake detection toolbox, which will be continuously updated to keep pace with advancements in deepfake technology.

–

May 28, 2024 – by Cass Kennedy