Bill C-36 Is Coming: What Canadian Organizations Using AI Need to Prepare Before the PPCDA Passes
Canada has been operating in an AI regulation vacuum for nearly two years. The Artificial Intelligence and Data Act — Part 3 of Bill C-27 — died on the Order Paper in January 2025 when Parliament was prorogued, and no equivalent AI law has been introduced since. The federal government's "AI for All" strategy, released on June 4, 2026, confirmed that AIDA would not be revived.
But the absence of a standalone AI statute does not mean the regulatory environment is stable. Bill C-36 — the Protecting Privacy and Consumer Data Act (PPCDA), tabled June 15, 2026 — introduces specific new obligations that apply directly to organizations using automated tools to make decisions about individuals. If your organization uses AI for credit assessment, employee performance scoring, customer segmentation with consequential outcomes, loan approvals, or any other process where an algorithm determines what a person receives or is denied, Bill C-36 creates new duties you need to understand now — before the bill becomes law.
The window is real but uncertain. Bill C-36 must pass through several parliamentary stages before it comes into force. That gap is preparation time, not permission to wait.
What Bill C-36 Is — and What It Dropped
Bill C-36 would repeal Part 1 of PIPEDA and replace it with the new PPCDA. It is not a re-run of Bill C-27. The Artificial Intelligence and Data Act framework has been dropped entirely. Canada's approach now handles AI governance through the privacy law rather than through a standalone AI statute.
That is a significant architectural choice. Rather than defining "high-impact AI systems" and imposing obligations based on AI risk categories — which was AIDA's approach — Bill C-36 imposes obligations on *any organization* that uses automated decision-making to reach decisions with a legal or similarly significant effect on individuals. McCarthy Tétrault's analysis of the bill notes that the definition of automated decision system is deliberately broad: rules-based systems, regression analysis, predictive analytics, machine learning, deep learning, and neural networks all fall within scope. If your system makes a consequential decision using data about a person, the PPCDA's explanation duties apply.
A companion bill, the Safe Social Media Act (Bill C-34, tabled June 10, 2026), proposes a Digital Safety Act enforced by the same regulator as the PPCDA. Neither has moved beyond first reading as of November 2026. Law firms reviewing the PPCDA consistently advise organizations to build rather than wait — both because the parliamentary calendar is uncertain and because the documentation and process work the bill requires takes months, not days.
Three AI-Specific Obligations the PPCDA Introduces
The PPCDA's AI-related requirements cluster around three areas that differ materially from current PIPEDA obligations.
Automated decision-making transparency. Under the PPCDA, organizations must inform individuals when an automated decision system makes a decision with a legal or similarly significant effect on them. That notification duty must happen proactively — not just be made available on request. When an individual asks for more detail, the organization must provide an explanation covering how the decision was reached, the data used, and the principal factors involved. Bennett Jones notes that affected individuals also gain the right to make written submissions to a human employee who can review the decision — a human-in-the-loop requirement that does not currently exist under PIPEDA.
This is new operational infrastructure, not just a documentation update. If your organization makes automated decisions about customers, employees, or applicants and cannot currently explain the factors driving those decisions in terms a non-technical person can understand, the PPCDA creates a legal obligation to build that capability before the law comes into force.
Enhanced consent for AI-processed data. The PPCDA raises the standard for meaningful consent. Organizations must disclose — in plain language — the purposes for which personal information will be processed, the methods used (including whether automated tools will be applied), the reasonably foreseeable consequences of that processing, the types of information involved, and the third parties who will receive it. For many organizations that deployed AI tools in the past two years without updating their privacy notices, this requires a full audit of what their tools do and a rewrite of consent language. Privacy notices that describe data use in generic terms — "to improve our services" — will not satisfy the PPCDA's disclosure standard.
Cross-border transfer obligations. The bill introduces an explicit requirement that organizations assess and mitigate privacy risks before transferring personal information outside Canada to any third-party processor. Gowling WLG's analysis of the PPCDA describes this as a meaningful tightening from PIPEDA's current contractual-accountability model. Under PIPEDA, comparable obligations often ran through vendor contracts. Under the PPCDA, the obligation is statutory and requires documented risk assessment — not just contract terms — before data is transferred. For cloud-hosted AI platforms, which describes virtually every major AI tool available, that assessment applies before the tool is used with personal information.
For Quebec-based organizations or any Canadian business operating in Quebec, Law 25's Privacy Impact Assessment requirement already applies to cross-border transfers. The PPCDA would extend a similar obligation to the rest of Canada, putting all organizations on a consistent standard.
The New Enforcement Regime
PIPEDA's enforcement model — a Privacy Commissioner who can make recommendations but cannot directly impose penalties — has long been criticized as toothless. Bill C-36 changes that. Miller Thomson's review of the PPCDA describes the enforcement upgrade as the most substantive change in Canadian federal privacy law since PIPEDA was enacted in 2000.
Oversight would move to a new regulator — the Digital Safety and Data Protection Commission of Canada — with direct order-making powers and the authority to impose administrative monetary penalties of up to $25 million or 5% of global annual gross revenues for knowing violations, whichever is higher. For mid-market companies, the percentage-of-revenue figure is the operative number. A business with $40 million in annual revenue facing a 5% penalty is looking at a $2 million exposure. That changes the risk calculus for organizations that have treated privacy compliance as a documentation exercise rather than an operational discipline.
The PPCDA also gives individuals the right to object to automated decision-making. Where an organization cannot demonstrate that its automated system is operating within the scope of valid consent and the bill's transparency requirements, a complaint to the new Commission is a plausible outcome.
The Readiness Gap Is Wider Than Most Organizations Expect
The Office of the Privacy Commissioner's 2025-26 survey of Canadian businesses found that 16% of Canadian businesses were using AI in their operations, up from 6% in 2023. That growth is real. The governance infrastructure to match it is not keeping pace. PwC Canada's 2026 Trust in AI Report found that 61% of Canadian businesses cite unclear or evolving legal and regulatory requirements as a major challenge for implementing AI responsibly.
"Unclear requirements" is a reasonable description of the current environment, in which AIDA died and no equivalent exists. Once the PPCDA passes, the requirements for automated decision-making transparency, consent, and cross-border transfers will be statutory obligations with real penalty exposure. Organizations that have used the regulatory gap to defer governance work will face a compressed preparation window.
The federal AI for All strategy targets raising Canadian AI adoption from roughly 12% to 60% by 2034. With only 8% of Canadian SMEs currently using AI, compared with 26% in Germany, the adoption curve is expected to accelerate materially. Organizations adopting AI tools before the PPCDA passes without building governance structures risk inheriting statutory compliance obligations they are not structured to meet on day one.
Four Things to Do in the Preparation Window
Bill C-36 is at first reading and must pass several more parliamentary stages before it becomes law. That window — which could range from several months to well over a year depending on the parliamentary calendar — is the time to build rather than wait. The organizations that enter the new regime ready are the ones doing the work now, not the ones scrambling at royal assent.
1. Inventory your automated decision systems. Map every process in your organization where an algorithm influences a decision about an individual customer, employee, or applicant. The PPCDA's definition is broad enough to include rules-based scoring systems and predictive analytics, not just large language models. The inventory is the foundation for every other step — without it, you cannot know which consent notices need updating, which data flows need assessment, or which decisions need an explanation capability.
2. Document data flows for cross-border transfers. Identify which of your AI tools process personal information on infrastructure outside Canada. For each, document what personal data flows there, under what contractual terms, and in which jurisdiction it is processed. This documentation serves dual purposes: it satisfies the PPCDA's transfer assessment requirement and it surfaces vendor contracts that may need renegotiation before the law comes into force.
3. Audit and update consent language. Review the privacy notices, terms of service, and consent flows that apply to data processed by your AI tools. Where the current language does not disclose that automated decision-making is used, the methods, or the reasonably foreseeable consequences, it needs to be updated before the PPCDA comes into force. For public-facing applications, this may require a re-consent process for existing users — work that takes time to plan, design, and execute.
4. Build your explanation capability for automated decisions. The PPCDA's explanation duty requires organizations to describe, to an affected individual, how an automated decision was reached and what data drove it. Many organizations that have deployed AI tools off the shelf do not have this capability — the tool produces an output but does not log the factors in a form that can be communicated to a human. Building that capability during deployment planning is substantially easier than retrofitting it after the fact.
None of these steps requires waiting for the bill to pass. All of them make your organization more defensible under the current PIPEDA framework while positioning you for the PPCDA.
Sources
- Prime Minister Carney. *AI for All: Canada's New National Artificial Intelligence Strategy.* pm.gc.ca (June 4, 2026)
- KPMG Canada. *Canada's National Artificial Intelligence Strategy: AI for All.* kpmg.com (June 2026)
- DLA Piper. *Canada tables Bill C-36: The Protecting Privacy and Consumer Data Act.* dlapiper.com (June 2026)
- Bennett Jones LLP. *Bill C-36 and Canadian Privacy Reform.* bennettjones.com (June 2026)
- McCarthy Tétrault. *Bill C-36: What Organizations Need to Know About Canada's Proposed Privacy Reform.* mccarthy.ca
- Gowling WLG. *Understanding Bill C-36.* gowlingwlg.com (2026)
- Miller Thomson LLP. *Bill C-36's PIPEDA-to-PPCDA Shift: The Rise of Documented Privacy Compliance.* millerthomson.com
- Office of the Privacy Commissioner of Canada. *2025-2026 Survey of Canadian Businesses on Privacy-Related Issues.* priv.gc.ca
- PwC Canada. *2026 Trust in AI Report.* pwc.com
- Tech Insider. *Canada AI Strategy 2026: $2B AI for All, 250K Jobs.* tech-insider.org
Cloud Forces' AI Advisory service includes structured preparation for incoming Canadian privacy legislation — from automated decision-making inventory to consent language review and cross-border data flow documentation. As an AWS Consulting Partner since 2019 and a Microsoft Solutions Partner since 2022, we bring the infrastructure depth to map your AI toolstack alongside the advisory expertise to build governance programs that meet the obligations the PPCDA will introduce. Book a consultation to begin your PPCDA preparation now.
Anton Kuznetsov is the founder and principal engineer of Cloud Forces, the Toronto firm he started in 2018 to make AI and cloud practical for Canadian SMEs. He leads Cloud Forces’ AI advisory and Claude deployment work and oversees the secure cloud platforms the firm runs for its clients.
Ready to bring AI to your business?
Book a free AI Readiness Consultation — no commitment required.
Book Free Consultation