Flipping the Script on Deepfake Technologies

Built for Change

An executive at an energy firm based in Britain receives a call from his boss, asking him to transfer 220,000 euros to a Hungarian vendor. He transfers the money, only to find out later that it wasn’t his actual boss who called him. It was a sophisticated fake of his boss’s real voice. You may have heard of “deepfakes” in the media by now. But what about the underlying technologies that power them, called GANs (Generative Adversarial Networks)? While most people probably associate deepfakes with scams or viral social media videos, there’s enormous potential for using GAN technology in legitimate business applications — from product design, to synthetic data, to art and content creation. But these new technologies also present an abundance of ethical, privacy and security problems. In this episode, we’ll speak with Edy Liongosari, Chief Research Scientist at Accenture Labs; Cristóbal Valenzuela, Co-Founder and C.E.O. at Runway; and David Danks, Professor of Data Science and Philosophy at the University of California, San Diego.