Deepfakes – synthetic images, audio and videos generated by artificial intelligence (AI) – have been proliferating at an incredible rate. In response, there have been several recent amendments to Canadian legislation to address the potential harms posed by this technology, including harms such as non-consensual distribution of sexually-explicit deepfakes and consumer and commercial fraud. The following is a round-up of recent amendments to Canadian legislation addressing deepfakes and a summary of some issues to watch out for.
1. Intimate Images Legislation
According to a recent Government of Canada report (“Report”), the vast majority of deepfakes available online (over 90 per cent) are non-consensual pornographic videos of women. Accordingly, the majority of Canadian legislation addressing deepfakes does so in the context of the non-consensual distribution of intimate images.
Criminal Law: Protecting Victims Act
On June 18, 2026, Bill C-16, the Protecting Victims Act, received royal assent. The Act seeks to address, among other things, the distribution of sexually explicit deepfakes. Under s. 162.1 of the Criminal Code, it is illegal to distribute an “intimate image” of a person without the consent of the person depicted. This provision was previously interpreted as not applying to deepfakes. For instance, in the March 2026 decision of R v MSK, 2026 NSPC 12, Judge Bronwyn Duffy of the Nova Scotia Provincial Court ruled that synthetic images did not meet the definition of an intimate image. Judge Duffy wrote at para. 105 that “[t]he law does not reflect the technology that exists today, and to combat advancing technologies in the commission of offences requires the dictates of Parliament”.
The Protecting Victims Act expands the offence to apply to sexual deepfakes by redefining “intimate images” to include “a visual representation that is made by any electronic or mechanical means, including by means of artificial intelligence software … if the depiction is likely to be mistaken for a visual recording of that person.” The Act also makes it illegal to threaten to distribute intimate images, including sexual deepfakes, without consent. The maximum penalty for non-consensual distribution, or threat of non-consensual distribution, has been increased from a maximum of 5 years imprisonment to 10 years.
Civil Liability for Non-Consensual Distribution of Sexual Deepfakes
All common law jurisdictions except Ontario and the Territories have enacted legislation creating liability for non-consensual distribution of intimate images, and in British Columbia, Manitoba, New Brunswick, Prince Edward Island and Saskatchewan, for threatening to distribute intimate images. The intimate images legislation in British Columbia, Saskatchewan, Manitoba, Quebec, New Brunswick and Prince Edward Island applies to AI-generated images and deepfakes, while Alberta, Nova Scotia and Newfoundland and Labrador’s legislation does not.
Manitoba’s legislation most explicitly addresses deepfakes among the common law jurisdictions. Under Manitoba’s The Non-Consensual Distribution of Intimate Images Act, an “intimate image” is defined as either a “personal intimate image” or “a fake intimate image”. A “fake intimate image” is defined as a sexually explicit visual recording that, in a reasonably convincing manner, falsely depicts an identifiable person, is created through the use of software, machine learning, artificial intelligence or other means, and in respect of which it is reasonable to suspect the person depicted would not consent to the recording being made or distributed. The Manitoba Act was amended effective June 1, 2026 to expand the orders a court can make, including requiring Internet service providers to remove intimate images from their online platforms.
In British Columbia, Saskatchewan, New Brunswick, and PEI, the definition of “intimate image” includes the phrase “whether or not the image has been altered in any way”, which has been interpreted as covering deepfakes. For instance, the guidance to British Columbia’s legislation provides that “[i]ntimate images can be real, fake or altered. This includes digitally altered images, digitally altered videos (deep-fakes) and AI-generated material.” In one 2025 decision of the British Columbia Civil Resolution Tribunal (A.Q. v. B.T., 2025 BCCRT 398), the Tribunal awarded damages to the applicant, a trans woman, for the non-consensual distribution by the respondent of an image showing the applicant’s head on a masculine body. Quebec’s Act to counter non-consensual sharing of intimate images, which came into force on December 4, 2024, defines an intimate image as “any image, altered or not, that represents or appears to represent a person” … where the person had a reasonable expectation that their privacy would be protected”.
Potential civil remedies for the non-consensual distribution of intimate images include a declaration, damages, an order to account for any profits that have accrued to the defendant, an injunction, an order to cease distribution, an order that an internet intermediary make reasonable efforts to cease distribution, and any other order that the court considers just and reasonable.
2. Consumer and Commercial Fraud
Another of the most common harms associated with deepfakes is consumer fraud. Fraudsters may use deepfakes of celebrities and other well-known public figures to sell fake products. The Report notes a well-known example of a deepfake of Elon Musk promoting fraudulent cryptocurrency giveaways, resulting in millions of dollars in financial losses. Businesses are not immune: another common scheme is the use of voice clones of CEOs and other senior representatives of high net-worth companies, which instruct staff to transfer money into the scammers’ bank accounts.
On June 12, 2026, Quebec’s Bill 24 – An Act to protect consumers against the misleading or fraudulent use of a person’s identity or image, came into force. Bill 24 amends Quebec’s Consumer Protection Act to prohibit the use of a person’s identity or image, without the person’s consent, for the purpose of making representations to a consumer. It also prohibits allowing a person’s identity or image to be used in such a context. “Image” is defined broadly to include any image, altered or not, that represents or appears to represent a person. It also includes any video or audio recording of that person.
A person whose identity has been stolen and a consumer who has been defrauded as a result can file a complaint with the Office de la protection du consommateur. Bill 24 grants new powers to the president of the Office de la protection du consommateur and the President and Chief Executive Officer of the Autorité des marchés financiers to issue orders requiring any person to cease engaging in, or to cease allowing anyone else to engage in, any practice involving the use of another person’s identity or image without that person’s consent. Contraventions of these new provisions may result in administrative penalties and penal sanctions.
The scope of liability under Bill 24 is broad, since it includes an organization that “allow[s]” the use of a person’s identity or image, without the person’s consent, for the purpose of making representations to a consumer. Organizations must therefore ensure that all the content that they produce or host does not breach these new provisions.
3. Bill C-277 (Regulating the Online Use of Deepfakes Act)
Bill C-277 – An Act to provide for the regulation of the online use of deepfakes and for related transparency measures, is a private member’s bill introduced by MP Michael Coteau that had its first reading on May 6, 2026. The bill would enact the Regulating the Online Use of Deepfakes Act to regulate the online use and dissemination of deepfakes, to increase transparency and to protect individuals and their interests from significant harm. It would also establish administrative monetary penalties to promote compliance with the Act. As a private members bill, Bill C-277 is unlikely to become law.
4. Deepfakes in the Employment Context
While no specific legislation has yet been passed, issues of deepfakes are beginning to crop up in the employment and labour context. In the union environment, a number of grievors, faced with video evidence of alleged wrongdoing, have called into question the bona fides of the video images, claiming that they have been altered with AI or are outright deepfakes (see, for instance, Paynter v. Deputy Head (Canada Border Services Agency), 2024 FPSLREB 116).
In another case (Rehn Enterprises Ltd. v United Steelworkers, Local 1-1937, 2024 CanLII 72130), the installation of video and audio surveillance cameras in company-owned vehicles was found by an arbitrator to be a violation of union members’ privacy interests in part because of the risk that the footage could be used to create deepfakes:
“While the Employer discounts the [employees’] concerns about “deep fakes” and AI, I find they are not fantastical but are grounded in legitimate worries about control over one’s personal information. As noted above, the rear-facing camera has been collecting and storing highly sensitive personal information since February 2023. While I have found that its continued collection is unreasonable, this does not minimize the concerns raised by the indefinite storage of the [employees’] images and voices to date or the importance of having effective policies in place to maintain the security and confidentiality of their personal information.”
5. Takeaway
Recent legislative changes addressing deepfakes are likely just the beginning, and organizations will have to keep pace with a rapidly changing AI regulatory landscape. Outside of legislation addressing the harms associated with deepfakes, other legislation addressing AI-related privacy concerns and personalized algorithmic pricing is also coming down the pipe.
Dentons’ privacy and cybersecurity group will be watching these developments closely. For more information on this subject, or for general advice regarding privacy and cybersecurity, please reach out to Kathryn Gullason or other members of the Dentons Privacy and Cybersecurity group.
