Deepfakes and Non-Consensual Pornography

Declarations: The Human Rights Podcast

The Deepfake detection platform Sensity came out with a report in 2019 that 96% of Deepfakes on the internet are pornographic and 90% of those represent women. Deepfakes are a modern form of synthetic media created by two competing AI’s with the goal of replicating hyper-realistic videos, images, and voices. Over the past five years this has led to major concerns of the technology being used to spread mis/disinformation, carry out fraudulent cybercrimes, tamper with human rights evidence, and most importantly in relation to this episode create non-consensual pornography. In this episode, the last of this season of the Declarations podcast, host Maryam Tanwir sat down with panellist Neema Jayasinghe and Henry Adjer who is not only responsible for the Sensity report that came out in 2019 but is also a seasoned expert on the topic of deepfakes and synthetic media. He is currently the head of policy and partnerships at Metaphysic.AI and also co-authored the report ‘Deeptrace: The State of Deepfakes’ while at Sensity. This was the first major report published to map the landscape of deepfakes and found that the overwhelming majority are used in pornography.