Enhancing Video Conferencing Security To Prevent Digital Injection Attacks

The Department of Homeland Security (DHS) is facing an increased threat to video conferencing communication from digital injection attacks. Digital injection attacks can be used to commit fraud and are able to bypass legacy tools designed to detect deceitful modification of image and video content. In response, Hive proposes to architect a solution that can verify the integrity of real-time video communication. Hive will prepare a detailed prototype plan that leverages its deep bench of technical expertise as a cutting-edge software developer along with its existing, best-in-class deepfake detection technology. The primary technical objective of the proposed Phase I research will be evaluating the feasibility of integrating Hive’s deepfake detection technology into existing video communication platforms like Zoom or Microsoft Teams to accurately identify digital injection attacks. This includes identifying the necessary capabilities and additional functions required for video applications, and specifying where these should reside within operating systems, device drivers, application software, and camera hardware. Hive will also explain how the solution seamlessly integrates with current systems and distinguish it from existing commercial presentation attack detection solutions. As part of the final deliverable, Hive will prepare an in-depth report that carefully outlines the required next steps to build a prototype that can successfully secure real-time communications. Hive expects that this research will lead to the development of a solution with wide-scale applicability — including within the DHS, other government agencies, and the private sector broadly.