Deepfakes! As more sophisticated, more personalized, more convincing audio and video manipulation emerges how do we get beyond apocalyptic discussion around malicious deepfakes and other AI-manipulated synthetic media and instead focus on what to do about malicious deepfakes and AI-manipulated synthetic media? Based on WITNESS’ collaborations with technologists, journalists, deepfakes researchers and human rights activists we'll explore state-of-the-art usages and types of deepfake and synthetic media, discuss the most relevant potential threat models and review the best potential technical and journalistic solutions for journalists and civic media. One particular area of focus will be the intersection of OSINT practice with deepfakes challenges.
The workshop will build on the ongoing focus of work that WITNESS has lead in this area including threat identification and modelling workshops with international and domestic participants at recent convenings, including the first cross-disciplinary expert convening on this area with First Draft in June 2018, workshops at MisinfoCon and StratComm, as well as an off-the-record senior journalists’ convening with First Draft, Knight Foundation, the Ethics and Governance of AI Initiative and a convening of platform representatives. We are also conducting workshops to connect key frontline verification and “fake news” debunkers from within the OSINT and journalistic community with the leading research community to test/learn one another’s methods, and discern what to adopt and how to make techniques accessible.
Organised in association with WITNESS.
