On May 21, 2026, Canada's four digital regulators — the Competition Bureau, the Canadian Radio-television and Telecommunications Commission (CRTC), the Office of the Privacy Commissioner of Canada (OPC), and the Copyright Board — convened an interactive Canadian Digital Regulators Forum (CDRF) workshop on artificial intelligence. Around the virtual table sat interim Commissioner of Competition Jeanne Pratt, CRTC Chair Vicky Eatrides, Privacy Commissioner Philippe Dufresne and Copyright Board Vice-Chair Drew Olsen, each examining how AI is reshaping their statutory turf and, more importantly, how their turfs increasingly overlap.
For a country whose tech-policy file is often parcelled out across silos, the workshop is a quietly significant moment. It is worth taking seriously — and pushing on.
The strongest case for the CDRF model
The instinct behind the CDRF — established in June 2023 and expanded to include the Copyright Board in September 2024 — is genuinely sound. AI systems do not respect regulatory boundaries: a foundation-model deal touches merger review, training-data consent, copyrighted inputs, and downstream broadcasting all at once. Four agencies issuing four uncoordinated demands would produce both compliance chaos and enforcement gaps.
The model is not a Canadian invention. The UK's Digital Regulation Cooperation Forum has been running since 2020, and similar arrangements now exist in the Netherlands, Ireland and Australia. A CRTC research paper on "collaborative oversight" notes that these forums "increase regulators' capacity for evolving digital regulations by bringing competition, privacy, and communications regulators together." That is exactly the cross-cutting capacity the AI moment demands.
The CDRF's first concrete output — a September 18, 2025 joint report on synthetic media — illustrated the upside. Rather than four separate regulator papers, the public got a single document mapping deepfake risks across copyright, broadcasting, competition and privacy, with consistent definitions and converging recommendations on labelling, consent and international alignment.
The substantive tests, in two recent cases
The coordination question is not academic. Two recent files show how the four regulators' AI work is already converging in practice.
First, the Competition Bureau's Artificial Intelligence and Competition: What We Heard report, published January 27, 2025, summarised 28 submissions to its 2024 AI consultation. The headline concern: a small cluster of incumbents — Amazon, Google, Microsoft, Meta and IBM — controls the compute, data and specialist talent that any serious AI entrant needs. Respondents flagged algorithmic-pricing collusion and deepfake-enabled deceptive marketing as the conduct risks most likely to outpace the Competition Act's existing tools.
Second, on May 6, 2026, the OPC and its Quebec, B.C. and Alberta counterparts released findings from a three-year joint investigation into OpenAI. They concluded that the initial ChatGPT model was trained in violation of Canadian privacy laws — "without adequate safeguards and valid consent," in the regulators' words — and that the company had relied on a "deceptive design pattern" tying chat-history access to training consent until April 2024. Crucially, the matter was resolved conditionally: OpenAI committed to filtering personal data from training sets, tighter retention rules, and ongoing transparency, and Commissioner Dufresne accepted that those measures "will address the concerns identified."
This is what proportionate enforcement looks like: a real finding of non-compliance, real remediation, no fine, no innovation-stopping order. It is also the template the CDRF should aspire to extend across competition, copyright and broadcasting.
Where the model gets risky
The danger is not the CDRF itself; it is what regulators do once they are coordinated. Four agencies pulling in the same direction can become four agencies stacking the same compliance load. A Canadian AI startup could plausibly face Bureau scrutiny of a compute-access partnership, an OPC consent inquiry into its training corpus, Copyright Board licensing tariff exposure, and CRTC questions about whether AI-assisted content qualifies as "Canadian." Each is defensible in isolation; the cumulative drag is what will determine whether Canada is a place ambitious AI companies build from, or one they avoid.
The CRTC's case is the one worth watching. Unions have already asked the regulator to exclude AI-generated work from Canadian-content credit. A blanket exclusion would punish the most likely productivity gain — cheaper dubbing, captioning and post-production for Quebec's roughly 750 dubbing artists and the broader sector — for a category error about authorship.
The surveillance side also matters. The Electronic Frontier Foundation's May 11, 2026 analysis of Bill C-22 warns that the legislation revives the most invasive lawful-access provisions of last year's failed Bill C-2 under a border-security framing. A digital-regulators forum that quietly normalises broad data-access powers — even for compelling AI-fraud or CSAM rationales — would undercut the OPC's own consent-and-transparency posture in the OpenAI matter.
What good looks like
The CDRF should treat the May 21 workshop as the start of three commitments worth making explicit:
- A single AI-compliance roadmap. Companies should be able to read one document that tells them, across all four regulators, what they must do — and just as importantly, what they will not be asked to do twice.
- Conditional-resolution as the default. The OpenAI outcome shows enforcement can be credible without being punitive. That posture should extend to merger review of AI deals and to copyright-licensing rate-setting.
- A presumption against extra-territorial overreach. Canadian rules should govern conduct in Canada and harm to Canadians; they should not become a parallel global AI regime.
Get that balance right and Canada's four-headed regulator becomes a genuine governance advantage. Get it wrong and it becomes four taxes on the same startup.