Very likely, in substance — but the CRTC page that would confirm the exact definition could not be reached this session, so treat the classification as the honest, cautious reading, not a verified quote.
Short answer
Probably, but we're flagging this one as partly unverified rather than guessing. Outbound calling in Canada is regulated under the Telecommunications Act, through rules the CRTC makes for unsolicited telecommunications — a separate regime from CASL, which governs electronic messages. That regime has historically targeted equipment that delivers a call using a recorded or synthesized voice, and a general-purpose AI voice agent making outbound calls fits that description in substance. The CRTC's own page setting out the current rule text was not reachable this session, so we're not quoting or reconstructing its exact wording — verify it directly before relying on any specific threshold or exemption.
The Telecommunications Act gives the CRTC its authority here: under s.41(1), “the Commission may, by order, prohibit or regulate the use… of the telecommunications facilities of a Canadian carrier for the provision of unsolicited telecommunications… to prevent undue inconvenience or nuisance.” Section 41(2) then draws a clean boundary: that power does not reach a message already covered by CASL s.6. In practice this means email and other electronic messages sit under CASL, while unsolicited phone calls — live, recorded, or synthesized — sit under the CRTC's own telemarketing and unsolicited-telecommunications rules, made under this section of the Act.
The CRTC's own site, which is where the precise definition of an automated-dialling device and the current rules for it actually live, returned no reachable page when checked for this piece — the domain did not respond with usable content. We are deliberately not repeating a specific definition or quote we could not check directly against a live page, because a wrong section number or a stale quote is invisible to a reader and worse than an honest gap. If your business runs outbound AI voice calls, confirm the current rule text directly against the CRTC's own published rules, or with counsel, before relying on any specific carve-out.
The Competition Bureau’s own discussion paper on AI and competition names exactly this pattern as a live concern: “AI eases the ability to scale such conduct, for example through automating phone calls or other telecommunications,” and flags that generative AI “could be leveraged to generate content for deceptive marketing conduct, such as e-mail or phone scripts,” including through synthesized voices. That is a federal competition regulator treating an AI-run outbound-calling programme as the kind of activity its deceptive-marketing rules already watch for — consistent with, even if not identical to, the CRTC's own telemarketing framework being read to cover it.
Until the classification is confirmed against a live CRTC source, the safer default is to treat outbound AI voice calls the way you would treat any automated telemarketing call: get actual express consent from the recipient before calling. Treadstone Law’s overview of Ontario advertising and marketing law covers the broader claims-and-conduct side of a marketing programme. For the email channel specifically, see whether AI email marketing is legal in Canada, and for consent mechanics generally, whether existing consent covers AI-sent messages.
See how AI fits into a compliant growth and marketing engine.
