Deepfakes
Produced in partnership with Kelsey Farish of DAC Beachcroft
Deepfakes

The following TMT practice note produced in partnership with Kelsey Farish of DAC Beachcroft provides comprehensive and up to date legal information covering:

  • Deepfakes
  • What is a deepfake?
  • How do deepfakes work?
  • Machine learning
  • Deep learning and GAN
  • Generation
  • Use of deepfakes
  • Applications of deepfakes
  • Entertainment
  • Parody and political satire
  • More...

This Practice Note considers the following legal issues, technical information and applications of deepfakes:

  1. What is a deepfake?

  2. How do deepfakes work?

  3. Applications of deepfakes

  4. Applicable UK law

  5. Corporate controls—what social media platforms and search engines are doing to counter

  6. Technological controls—future tech counters to deepfakes

  7. Regulation and control—problem areas concerning technological controls

  8. Regulation and control—problem areas concerning legislative or common law controls

What is a deepfake?

The term deepfake is a portmanteau of ‘deep learning’ and ‘fake’. A deepfake is a form of synthetic media in which existing media, for example, a video clip and a series of photographs, are combined together using a sophisticated type of artificial intelligence (AI) to create a realistic but fake video. Since the technology’s inception in the mid-2010s, ‘deepfake’ has come to refer to any face-swapping technique whereby images of an individual are used by AI to generate digital doppelgängers (look-alikes) and then superimposed onto different bodies. Even manipulated videos which do not utilise AI technology are commonly referred to as deepfakes. By late 2020, the vast majority of detected deepfakes featured individuals in the entertainment, fashion, or sports sectors with major players such as Disney actively developing their own variants.

How do deepfakes work?

Computers have been used to generate or otherwise animate the human face since the 1970s. By the early 2000s, the technology had dramatically improved in terms of creating realistic, computer-generated

Popular documents