What if deepfakes made us doubt everything we see and hear?

At a Glance 07-09-2021

Deepfakes are hyper-realistic media products created through artificial intelligence (AI) techniques that manipulate how people look and the things that they appear to say or do. They hit the headlines in 2018 with a deepfake video of Barack Obama, which was designed to raise awareness of their challenges. The accessibility and outputs of deepfake generation tools are improving rapidly, and their use is increasing exponentially. A wide range of malicious uses have been identified, including fraud, extortion and political disinformation. The impacts of such misuse can be financial, psychological and reputational. However, the most widespread use so far has been the production of non-consensual pornographic videos, with negative impacts that overwhelmingly affect women. Deepfakes may also contribute to worrying trends in our media, as well as in our social and democratic systems. While the technology itself is legal, some malicious uses are not, and a combination of legal and technical measures may be mobilised to limit their production and dissemination.