DeepID: Detection Platform to Stop Adversary DeepFake Attacks

Technical Abstract: Russia is actively using synthetic media - “Deepfakes” - as an information warfare tool to support its invasion of Ukraine. A DeepFake is an AI generated or manipulated video, audio, or news-article that appears 100% realistic to the human eye and can fool existing biometric detection methods. The DAF/DoD lacks linguists in target threat languages to accurately analyze and detect deepfakes at scale. This allows Russian information warfare to go unstopped and gives Moscow an advantage in the war. DeepMedia proposed to modify, adapt, and enhance its deepfake detection platform, DeepIdentify, to enable the Air Force to identify adversary information war efforts through publicly available information sources faster, more accurately, and at scale. DeepIdentify is a deepfake detection platform that detects manipulated video, audio, and text produced by adversaries. DeepMedia will conduct a R/R&D feasibility study to adapt the platform to Air Force / DoD uses so the Air Force can identify, detect, and attribute manufactured deepfake content in order to counter enemy information warfare.