Not by law, for most businesses — but a voluntary code and ordinary advertising rules both reach further than that sounds.
Short answer
No. Canada has no statute requiring a business to label AI-generated content. The nearest national standard is a voluntary code that binds only the organizations that signed it, and even those obligations are narrower than most people assume. Separately, existing advertising law can still catch you if leaving a label off makes an ad misleading.
ISED’s Voluntary Code of Conduct commits signatories to six outcomes — accountability, safety, fairness and equity, transparency, human oversight, and validity and robustness. Within that, one specific line is the closest thing Canadian federal policy has to a labelling rule: a manager of a public-facing system must “ensure that systems that could be mistaken for humans are clearly and prominently identified as AI systems.”
Two words matter more than the rest of the sentence: it’s voluntary, and it binds signatories. The code itself says plainly that it “does not in any way change existing legal obligations that organizations may have” under Canadian law. As of this writing 46 organizations have signed on — if yours hasn’t, this specific obligation doesn’t apply to you at all.
Nothing in the voluntary code creates a general labelling law, but the general prohibition on deceptive marketing predates AI and doesn’t care how content was made. Ontario’s advertising rules explain the underlying principle: Canada’s Competition Act “prohibits misleading advertising and deceptive marketing practices across the country,” and that rule addresses the general legal principle of what counts as a misleading representation — it isn’t an AI rule at all, and applies whether the misleading element is a fake discount, a staged photo, or an unlabelled AI voice.
The Competition Bureau’s own discussion paper flags exactly the kind of content this could reach: it names “deepfake’ (digitally altered) voices, images, or videos” as a way AI could be “leveraged in deceptive marketing practices,” and notes plainly that “AI eases the ability to scale such conduct.” If an unlabelled AI element makes an ad deceptive to a reasonable consumer, the general rule can apply regardless of any voluntary code.
Label a chatbot or voice agent that could be mistaken for a person — it’s good practice whether or not you’ve signed the code, and it’s the specific behaviour the code’s signatories have already committed to. For an ordinary blog post, product photo, or marketing image, there is no general Canadian obligation to disclose AI involvement, so long as the content itself doesn’t make a false claim.
The image-specific version of this question — whether you can use AI images on your website — and the deepfake question, is making a deepfake illegal in Canada, both build on the same distinction between a voluntary code and enforceable law.
See how compliant, disclosed AI marketing actually gets built.
