Deepfakes, Counterfeits, and Personality

Deepfakes are synthetic audiovisual content generated by machine learning that can be used to simulate highly convincing representations of real people doing and saying things they have not done or said, consequently creating both private and public harms. Deepfakes are one tool among many to generate and spread online misinformation, and, especially when they depict prominent and widely recognizable people, can erode trust in political and cultural institutions. At the same time, deepfakes of individuals who are not public figures can cause serious damage to the autonomy, dignity, and integrity of those falsely depicted, with significant repercussions for their personal and professional reputations. These deepfake injuries to non-famous people may be expected to increase as the software to create deepfakes becomes more mainstream, more affordable, and produces more realistic results.

Although the digital platforms have implemented deepfake policies, they primarily address the public harms from misinformation and focus on mitigating hermeneutic harms from audience misperception. The platform’s deepfake policies only indirectly address the private harm experienced by individuals who are depicted in deepfakes without their consent, and only indirectly act to deter or penalize those who generate and circulate deepfakes. Hence, there is a complementary role for legal regulation of deepfakes to address these private harms.

We explain that the core private harm injury arising from the creation and circulation of a deepfake is a counterfeiting of personality. The non-consensual use in deepfakes of a person’s image, mannerisms, voice, and other traits results in a loss of autonomy and dignity relating to the undetected doubling of a person’s identity. We examine the existing personality-offenses in law, including personation in the Criminal Code, impersonation in elections law, and the tort of misappropriation of personality, and explain that they are designed to remedy these counterfeiting harms incurred by the undetected doubling of a person’s identity. We argue these personality-based offenses should be applied to penalize and deter those who generate and circulate deepfakes and to redress the harms arising from the counterfeiting of personality caused by deepfakes.