Artificial Intelligence and Data Act (AIDA) — Canada (Bill C-27, Part 3)

The Artificial Intelligence and Data Act (AIDA) was Canada's proposed horizontal AI statute, introduced in June 2022 as Part 3 of Bill C-27, the Digital Charter Implementation Act, 2022. It would have regulated "high-impact" and general-purpose AI systems used in interprovincial and international trade, supervised by an AI and Data Commissioner inside the Ministry of Innovation, Science and Industry. The bill stalled in committee and died when Parliament was prorogued and dissolved in January 2025, never reaching report stage or third reading. As of April 2026 no successor AI bill has been tabled under the Carney government.

Bill C-27 also contained the Consumer Privacy Protection Act (Part 1) and the Personal Information and Data Protection Tribunal Act (Part 2). It was proposed by the Government of Canada under the Liberal government of Prime Minister Justin Trudeau and tabled by the Minister of Innovation, Science and Industry, François-Philippe Champagne.

Status and legislative history

AIDA was introduced at first reading on 16 June 2022. It passed second reading and was referred to the House Standing Committee on Industry and Technology (INDU) on 24 April 2023. In November 2023 the government tabled amendments that significantly revised the high-impact-AI concept and added general-purpose AI (GPAI) provisions after industry and civil-society feedback. Committee hearings ran from September 2023 to September 2024, spanning more than 36 meetings.

Parliament was prorogued on 6 January 2025, killing the bill on the order paper before it reached report stage or third reading, ahead of the April 2025 federal election. As of April 2026 the bill had not been reintroduced. The Carney government — the Liberal minority re-elected in April 2025, with Mark Carney having succeeded Trudeau as prime minister — had not tabled a successor AI bill as of this page's last update.

Scope and definitions

AIDA would have been Canada's first horizontal AI statute. It applied to private-sector actors engaged in regulated activity in the course of international or interprovincial trade and commerce, using the federal trade and commerce power as its constitutional hook. Federal government institutions (governed separately by the Treasury Board's Directive on Automated Decision-Making) and national-security activities were excluded.

The central regulatory category was the "high-impact" AI system. This term was undefined in the original 2022 text and drew criticism on that basis; the November 2023 amendments added a list of seven classes: employment, service provision, biometric use, content moderation and prioritisation, healthcare, judicial and administrative decision-making, and law enforcement. The November 2023 amendments also added general-purpose AI systems (GPAI) as a category, tracking the EU AI Act (Regulation 2024/1689) GPAI tier, and placed separate obligations on developers of machine-learning models intended for high-impact use.

Key provisions

The provisions below reflect the November 2023 amended text.

High-impact system obligations

Developers, deployers, and operators of high-impact AI systems would have been required to:

• Identify and mitigate risks of harm and biased output
• Monitor compliance with mitigation measures
• Publish a plain-language description of the system
• Keep records of how data is anonymised, of the system's intended use, and of risk assessments
• Notify the Minister of material harms

GPAI obligations

The GPAI obligations, added in 2023, tracked EU AI Act logic. Providers would have been required to:

• Assess and mitigate risks of harm and biased output before making the GPAI system available
• Maintain technical documentation
• Publish information sufficient for downstream deployers to understand capabilities and limitations

AI and Data Commissioner

Supervision would have rested with an AI and Data Commissioner, a senior official designated by the Minister of Innovation, Science and Industry and housed within the department rather than constituted as an independent regulator. This was among the bill's most contested design choices.

Delegated regulation

Most operational detail — including what exactly qualified as "high-impact," the content of record-keeping, risk-assessment methodology, and penalty amounts — was to be set by Governor-in-Council regulations rather than in the statute itself. This was among the bill's most criticised features.

Enforcement and penalties

AIDA provided for administrative monetary penalties, with regulations to be made after enactment. It created new criminal offences for making AI systems available knowing they are likely to cause serious harm and for using unlawfully obtained personal data in AI systems, carrying fines up to $25M CAD or 5% of global gross revenue. A private right of action was not included.

Reasons for stalling

Analyses of why AIDA lost support cite several factors (Source: BABL AI analysis, White & Case Global AI Tracker, Chambers Canada Trends and Developments 2025):

• Delegated-regulation criticism. The parliamentary committee, industry, and civil society all complained that too much substance was pushed into future regulations, leaving the statute as a skeleton. This was the most-cited reason AIDA lost support.
• Independence of the AI Commissioner. Critics argued the commissioner should be a free-standing officer of Parliament rather than an official reporting to the Minister.
• Scope ambiguity. The initial bill did not define "high-impact"; the November 2023 amendments added a list but satisfied neither those who wanted broader coverage nor those who wanted narrower, sector-specific regulation.
• Bundling with privacy reform. AIDA shared Bill C-27 with contentious privacy-law provisions, making the package too ambitious to move through committee in a single Parliament.
• Political timing. The Trudeau government's declining support through 2024 and his December 2024 resignation announcement left Bill C-27 without sufficient political support.

Tensions and critiques

Several structural critiques recurred in the parliamentary and scholarly debate:

• Skeleton-statute critique. Critics argued that a law delegating "high-impact" definition, risk-assessment methodology, and penalty amounts to future regulation was vulnerable to being weakened or stalled at the regulation stage. This was the most-cited reason the bill lost support.
• Regulator independence. Critics including Ada Lovelace-adjacent Canadian scholars and the Canadian Civil Liberties Association argued that a minister-appointed commissioner could not credibly supervise AI deployment by federal-government contractors and politically salient services.
• Horizontal versus sectoral regulation. AIDA's horizontal approach overlapped with Canada's existing sectoral regulators (OSFI for finance, Health Canada for medical devices, the Privacy Commissioner, the Competition Bureau). Industry argued that sector-specific rules would be more appropriate; civil society countered that sectoral gaps would leave some AI uses uncovered.
• Federal-provincial friction. Parts of AI regulation — employment, consumer protection, healthcare — fall within provincial jurisdiction. Quebec's Law 25 (2024) and AIDA overlapped, and reconciliation was not fully resolved in the bill.
• Timing relative to the EU. By 2024 the EU AI Act (Regulation 2024/1689) had set the default global template, and AIDA's amendments increasingly tracked EU structure (the GPAI tier). Critics argued Canada should either adopt EU rules wholesale for interoperability or design a genuinely distinctive framework.
• The AISI gap. Unlike the UK, US, EU, Japan, and South Korea, Canada's Canadian AI Safety Institute (CAISI), announced in 2024, has no statutory footing. AIDA would not have changed that. The open successor question is whether Canada follows the UK DUAA model (data-protection reform plus a voluntary AISI) or the Korean and EU model (a comprehensive statute).

Carney government posture

This section carries medium-low confidence given limited primary-source reporting. Mark Carney took office as prime minister in March 2025 and won a minority mandate in April 2025. The Carney government's public AI posture has emphasised AI competitiveness, data-centre investment, and alignment with G7 partners rather than comprehensive horizontal regulation. A March 2026 joint Carney-Albanese statement on AI cooperation did not commit to a specific domestic AI bill. No successor AI bill had been tabled as of this page's last update. Industry groups including CanadaAI and the Council of Canadian Innovators have lobbied for a narrower, sector-specific approach rather than an AIDA-style horizontal statute.

In the absence of a federal AI statute, Canada's de facto AI regime rests on the 2023 Voluntary Code of Conduct on Advanced Generative AI Systems — signed by Cohere, OpenAI, Telus, BlackBerry, and others — together with the Treasury Board Directive on Automated Decision-Making for federal agencies. The lack of domestic guidance leaves Canadian-headquartered AI companies, Cohere in particular, choosing among EU, UK, or US compliance baselines.

Provincial action

On 29 April 2026, Manitoba Premier Wab Kinew announced that the province would introduce legislation banning underage users from AI chatbots and social media platforms (Source: cbc.ca). It was the first Canadian sub-national chatbot and minor ban, and a provincial-level counterpoint to the absence of federal AI legislation. The measure is functionally comparable to CA SB 243 and the US Senate CHATBOT Act, which advanced on the same day, with three independent jurisdictions in the same week converging on age-gated companion-AI restrictions. Provincial legislation can move without resolving the federal-provincial jurisdictional questions that constrained AIDA.

Comparison with other approaches

AIDA was Canada's first attempt at horizontal AI legislation and the first such attempt to fail among G7 democracies.

Relationships

• related: EU AI Act (Regulation 2024/1689) — the regulatory model AIDA's November 2023 amendments increasingly tracked (GPAI tier)
• related: Data (Use and Access) Act 2025 — Source Summary — parallel dynamic: AI-specific legislation deferred in favour of data-protection reform plus a voluntary AISI
• related: South Korea AI Basic Act — Source Summary — contrast: Korea enacted comparable horizontal legislation in Dec 2024, the same month AIDA effectively died
• related: Japan AI Promotion Act — Source Summary — contrast: pro-innovation statutory alternative
• related: Australia Voluntary AI Safety Standard — Source Summary — another voluntary-first jurisdiction with mandatory guardrails under consideration
• related: AI Sovereignty — AIDA's failure as a data point on mid-size democracies' capacity to build independent AI regulatory institutions
• related: Techno-Federalism: How Regulatory Fragmentation Shapes the U.S.-China AI Race — Canada's federal-provincial AI jurisdiction overlap

Sources

• AIDA source summary
• LEGISinfo (Parliament of Canada): https://www.parl.ca/legisinfo/en/bill/44-1/c-27
• ISED AIDA Companion Document (Mar 2023): https://ised-isde.canada.ca/site/innovation-better-canada/en/artificial-intelligence-and-data-act-aida-companion-document
• BABL AI, "Canadian AI Bill Stalls as Bill C-27 Terminates in Parliament" (2025)
• White & Case, AI Watch: Global Regulatory Tracker — Canada
• Chambers AI 2025: Canada Trends and Developments