Saskatchewan introduces bill targeting deceptive AI deepfakes
Bill 76 would make deceptive synthetic media a civil wrong, require labels on realistic AI content and let courts order removal, but it is not law yet.

Key takeaways
- Saskatchewan Justice Minister and Attorney General Tim McLeod introduced Bill 76, The Prevention of Deceptive Artificial Intelligence Act, for first reading on November 17, 2026.
- The bill defines "deceptive use" as synthetic media used to make someone reasonably believe content is authentic, where that belief is reasonably likely to cause substantial harm or a significant risk to public safety.
- It would make deceptive use a civil wrong that a plaintiff could bring without proving damage, and let courts award damages and order content removed or de-indexed, but it is proposed law that needs to pass and be brought into force by order of the Lieutenant Governor in Council.
- The government says Saskatchewan is the first Canadian jurisdiction with such a framework; Alberta has had a narrower political-deepfake law since April 2026.
Saskatchewan's justice minister introduced legislation on November 17, 2026 that would make the deceptive use of AI-generated "synthetic media" a civil wrong and require labels on realistic fake content. Bill No. 76, The Prevention of Deceptive Artificial Intelligence Act, was introduced for first reading by Tim McLeod, the Minister of Justice and Attorney General. The bill must pass the Assembly to become law and would take effect by order of the Lieutenant Governor in Council. None of its powers exist yet, and it states no start date and no compensation figures.
The bill defines synthetic media as audio, visual or digital content generated, synthesized or materially altered, including by AI, that imitates a person's image, likeness, voice or identity, or an event or conduct. "Deceptive use" is creating, publishing, distributing or using that content intending to make someone reasonably believe it is authentic, where that belief is reasonably likely to cause substantial physical, psychological, reputational or economic harm, or a significant risk to public safety. Section 3 would prohibit knowing deceptive use and leaves lawful uses, including parody and satire, untouched.
Section 4 would require realistic synthetic media to carry a clear and prominent disclosure and, where technically feasible, a persistent machine-readable marking, unless its artificial nature would be obvious to a reasonable person. Section 7 makes deceptive use a tort that a plaintiff could bring without proving damage, and section 8 would let a court award damages and order content removed from platforms or de-indexed from search engines. Related amendments would label synthetic media in election ads, with a parody or satire exception, add synthetic-media false endorsements to consumer-protection law, and fold synthetic images into the Privacy Act's intimate-image definition.
McLeod told reporters the bill does not bar lawful synthetic media, including parody or satire, and that compensation would depend on the harm caused, CBC reported. He cited a Nova Scotia case in which a judge acquitted a man charged under the Criminal Code's intimate-image offense over AI-generated images of classmates. The government's release says Saskatchewan becomes "the first Canadian jurisdiction to establish a legislative framework" for deceptive AI, a claim The Canadian Press repeats. Alberta's Justice Statutes Amendment Act, 2026, given royal assent in April 2026, covers political deepfakes only, a narrower scope.
Canadian law already reaches part of this. The Criminal Code's intimate-image offense now covers a visual representation made by electronic or mechanical means, including by AI software, that shows an identifiable person depicted as nude, as nearly nude, exposing their sexual organs or engaged in explicit sexual activity, where the depiction is likely to be mistaken for a visual recording of that person. Elections Canada told a Senate committee the existing impersonation offense does not capture voice or image manipulation. Parliament has since closed part of that gap: Bill C-25, the Strong and Free Elections Act, received Royal Assent on June 18, 2026, as Chapter 20 of the Statutes of Canada, 2026, and it bans sophisticated deepfakes of electoral actors that intend to mislead.
In the US, the federal TAKE IT DOWN Act requires covered platforms to remove nonconsensual intimate depictions, including AI "digital forgeries," within 48 hours of a valid request. Tennessee's 2024 ELVIS Act covers unauthorized voice and likeness, and California's election-deepfake law AB 2839 was permanently enjoined in 2025, according to a compliance guide tracking the cases.
For a creator making synthetic video for a Canadian audience, nothing changes unless the bill passes. If it does, the likely duties would be a visible label and, where feasible, machine-readable marking on realistic synthetic video and audio that could be mistaken for real, with the parody and satire carve-out intact.
Bill 76 proceeds to second reading; no date is set, and it takes effect only by order in council.
Sources
- saskatchewan.ca - government release, McLeod quote, "first Canadian jurisdiction" claim
- docs.legassembly.sk.ca - full bill text: definitions, prohibition, disclosure, tort, remedies, amendments, coming into force
- docs.legassembly.sk.ca - Bill No. 76 first reading by McLeod, Nov 17 2026
- cbc.ca - McLeod remarks on parody, compensation, Nova Scotia case
- globalnews.ca - Canadian Press report; Alberta comparison; "first Canadian jurisdiction"
- alberta.ca - Alberta Bill 23 royal assent April 16 2026, political deepfake prohibition
- laws-lois.justice.gc.ca - Criminal Code intimate-image offense now covers AI-made representations
- elections.ca - Elections Canada on the impersonation gap and Bill C-25
- ftc.gov - TAKE IT DOWN Act removal duty, digital forgeries, 48 hours
- superintelligencecompliance.ai - California AB 2839 injunction and Tennessee ELVIS Act, attributed