The Science of Audio Deepfake Detection

Deepfakes are becoming a widely deployed technology for both fun and profit. Within this space, audio deepfakes are particularly interesting as they have the capacity to improve people’s lives, such as giving someone back their voice when they have lost the ability to speak. However, this technology can also be used in advanced social engineering attacks and to perpetrate fraud. In this talk, we will provide an overview of how deepfakes are generated, along with how current detection technologies work. Along the journey of designing appropriate detection mechanisms, I’ll point out areas where we need more science and better application of the scientific method.

Bio: Dr. Carrie Gates

Dr. Carrie Gates joined Bank of America in October 2018 as a Senior Vice President in Global Information Security. She has established a research program, working in partnership with universities collaborators to pursue longer term, higher risk research projects in the security space that have the potential to improve the bank’s security posture. Her current portfolio includes the detection of audio deep fakes, mis- and disinformation, adversarial machine learning, and usable security.

Prior to joining the bank, Carrie worked for both Dell and CA Technologies as a Distinguished Engineer, and with CERT at Carnegie Mellon University. Prior to starting her research career, she was the systems manager for the Faculty of Computer Science at Dalhousie University (Canada). She has over 50 peer-reviewed publications and over 20 awarded patents in the computer and network security field.

Outside of work, Carrie enjoys riding horses, traveling and photography.