As AI continues to advance, concerns about the protection of personality rights, including the image and voice of individuals, have become increasingly urgent. Modern AI systems are now capable of producing outputs which closely mimic the personal characteristics of real people. While these technological developments are impressive, they are also highly alarming, especially given the limited ability of existing legal frameworks to offer adequate protection. AI systems rely on a significant amount of training data to function effectively, which raises critical questions about where this data is coming from and whether individual rights are being respected in its collection and use. Without clear legal boundaries, the use of personal characteristics in both AI training and AI-generated outputs risks diminishing individual autonomy over one’s identity.
Gaps in Canadian Law
In Canada, legal protection against the misuse of personality remains fragmented. Our system does not recognize a standalone right to personality, causing those seeking recourse to rely primarily on common law remedies.
One available option is the tort of misappropriation of personality; however, this action is often ill-suited to address the harm caused by AI. To succeed in a claim for misappropriation of personality, a plaintiff must show that the defendant used their name, image, or likeness, without consent, for commercial gain.[i] Meeting the requirements for this cause of action can be challenging, particularly in non-commercial contexts where someone’s image is replicated without direct profit.
Since the Canadian Copyright Act extends protection only to human-authored works, it remains unclear whether a human face can be understood as an “original creation” for the purposes of copyright law. If personality rights are to be meaningfully protected against the use of AI, Canada needs a clearer legal framework that directly addresses permissions and the scope of individual protections.
International Approaches
Other jurisdictions have begun to respond more directly to the challenges AI poses to personality rights. China, for example, recognizes a standalone right to personality, including voice, through its Civil Code.[ii] In a 2024 decision, the Beijing Internet Court ruled that the voice rights of an individual extend to AI-generated voices, emphasizing that clear consent is required before anyone’s voice can be used in the training of an AI system which, in this case, was for a text-to-speech application.[iii]
France relies on its existing legal framework to protect image rights, permitting legal action against non-consensual uses, particularly where the image is altered or harmful to an individual’s dignity.[iv]
Denmark has also taken significant steps in this area. In June 2025, Denmark announced proposed amendments to its Copyright Act, aimed at addressing the rise of AI generated “deepfakes” – realistic representations of a person.[v] These amendments would grant individuals rights over their body, facial features, and voice, allowing them to demand the removal of AI-generated content that was shared without their consent. These changes are expected to come into force this year.
The Path Forward for Canada
Canada may benefit from following Denmark’s lead by expanding copyright protections to cover personal rights. While this would be a significant shift in Canadian copyright law, it could offer individuals greater control over their identity in an increasingly digital environment.
Without targeted reform, Canadian law risks falling behind as AI technologies continue to blur the lines between human identity and software imitation.
[i] Gould Estate v Stoddart Publishing Co, 1996 ONSC 8209.
[ii] Civil Code of the People’s Republic of China, adopted 28 May 2020, art 1023.
[iii] Yin v Company A and Others, (2024) Beijing Internet Court (China).
[iv] Government of France, “Protect your image rights” (28 February 2024), art 226-8 of the Criminal Code.
[v] Miranda Bryant, “Denmark to tackle deepfakes by giving people copyright to their own features”, The Guardian (27 June 2025).
Any article or other information or content expressed or made available in this Section is that of the respective author(s) and not of the OBA.