Deepfakes as a growing threat

Deepfakes are images and videos created using artificial intelligence to mimic real people. Many are used to make non-consensual pornographic content, which can cause serious harm to individuals whose identities are stolen or manipulated.

Impact on victims

The harm from deepfakes goes beyond spreading false information. Victims face real-life consequences, such as emotional distress and damage to their reputation. Their identity and personal information are used against them without their consent, making the issue more than just a digital concern.

Real-world examples

In 2018, journalist Rana Ayyub faced cyberbullying using deepfake videos of her. The United Nations condemned the incident, highlighting how such content can be used to silence individuals. The harm is not only psychological but also affects a person’s sense of self and trust in others.

Deepfake applications beyond pornography

Deepfakes are used in various malicious ways, including political disinformation and fraud. In Slovakia, a fake audio clip influenced an election, while in Hong Kong, a worker was fired after a deepfake video call. These cases show how the technology can be weaponized in different contexts.

The concept of the digital body

The term "digital body" refers to the collection of signals that identify a person, such as their face, voice, and gestures. These signals form part of a person’s digital identity. When deepfakes use these signals without consent, it feels like an attack on the person themselves.

Legal and technical challenges

Current laws are not well-equipped to handle the complexities of deepfakes. Legal protections are fragmented, focusing on property or privacy rather than the integrity of digital identity. Researchers argue that simply detecting fakes does not help victims, as the harm is real and ongoing.

Historical context of human synthesis

Human synthesis, the digital creation of a person’s appearance and voice, has a long history. Early experiments in digitizing images date back to 1957, and the use of virtual actors in films like The Crow in 1994 shows how the technology has evolved over time.

The rise of the term "deepfake"

The term "deepfake" gained notoriety in 2017 when a Reddit user shared non-consensual videos of celebrities using AI. This led to the development of generative adversarial networks and other models that now dominate image creation. The term highlights a clear violation of consent.

Efforts to protect digital identity

Some countries are exploring new legal frameworks to protect digital identity. Denmark, for example, is considering granting individuals a direct right to their face and voice. This would make the use of someone’s image without consent illegal, regardless of intent or harm.

Technological solutions for prevention

Research projects like BeepFake aim to prevent deepfakes by making it harder to clone faces and voices. These efforts focus on protecting digital identities before content is created, ensuring that consent is the foundation of any use of personal data.