AI-Cafe presents: Deepfakes – An emerging Internet Threat and their Detection

Deepfakes have emerged for some time now as one of the largest Internet threats, and even though their primary use so far has been the creation of pornographic content, the risk of them being abused for disinformation purposes is growing by the day. Deepfake creation approaches and tools are continuously improving in terms of result quality and ease of use by non-experts, and accordingly, the amount of deepfake content on the Internet is quickly growing. For that reason, approaches for deepfake detection are a valuable tool for media companies, social media platforms and ultimately citizens to help them tell authentic from deepfake generated content. In this presentation, I will be presenting a short overview of the developments in the field of deepfake detection, and present our lessons learned from working on the problem in the context of the Deepfake Detection Challenge and from developing a service for the H2020 WeVerify project.