Synthetic media is the umbrella term for any image, video, audio or text that has been generated or substantially altered by AI rather than captured or written directly by a person — a category that covers everything from a wholly AI-generated marketing image to a voice-cloned narration, and includes a deepfake as one particular, deceptively-used kind of it.
Content Credentials’ own framing of the problem it exists to address uses this exact vocabulary: “Almost anyone can create realistic photos, videos, audio files, and documents with easy-to-use tools and a few clicks. While there will continue to be bad actors who seek to label synthetic content as authentic, our goal is to provide good actors with a way to demonstrate the authenticity of their content”. That sentence draws the line worth holding onto: synthetic media is not itself wrongdoing. A retailer’s AI-generated product photo and a fabricated video of a public figure are both synthetic media; only the second is commonly called a deepfake, and the label tracks how the content is used, not how it was made.
Canada has no statute that defines “synthetic media” or regulates it as a category. Where synthetic content becomes a legal problem, existing law supplies the hook — the Competition Bureau’s discussion paper on AI and competition treats a synthetic voice or image used in deceptive marketing as an aggravating detail within its existing deceptive-marketing analysis, not as a separately regulated technology. The same discussion-paper reasoning that applies to a “deepfake” voice used to make a scam call, in other words, applies just as much to any other synthetic recording used the same way.
A brokerage produces three pieces of content in one week: an AI-narrated market-update video using a licensed synthetic voice, an AI-generated banner image for a listing, and an AI-cloned recording of one broker’s own voice, made without telling clients, used to leave automated follow-up voicemails that sound like the broker personally called. All three are synthetic media. The first two are disclosed, licensed uses with no one misled about who or what produced them. The third borrows a real person’s identity to make clients believe a human placed the call — the point at which synthetic media shades into the deepfake pattern the Competition Bureau and deceptive-marketing law are actually aimed at.
See also: what is a deepfake, what is provenance metadata, is making a deepfake illegal in Canada.
Using synthetic voice, image or video in customer-facing marketing raises disclosure and brand-voice questions — ai-growth-marketing covers how to use it without misleading anyone.
