OpenMedia’s Input to Canada’s 2026 AI Transparency Public Consultation

Our relationships with reality are increasingly distorted, if Canada permits corporations unchecked powers over AI, the consequences can be destructive. OpenMedia offers extensive community-informed recommendations on AI transparency.

On September 23, 2026, OpenMedia submitted our written comments to the 2026 “Have your say on advancing AI transparency in Canada” public consultation. This two-month consultation aimed at advancing the vision in Canada’s national “AI for All” strategy on artificial intelligence, and focused on five major areas related to AI transparency specifically: AI-generated content, human interaction with AI systems, information about AI systems, AI incidents, and agentic AI.

In our submission, we stress the importance of accountability to human beings as the necessary foundation of AI transparency, needed to support innovation, rights, privacy and sustainability alike. As hyperscalers are gaining more unchecked power over technology, and our relationship with our society and environment are increasingly distorted by AI related activity, we welcome the Canadian government introducing this participatory process to ensure Canada understands the problem correctly with more diverse stakeholders before taking further regulatory steps.

Our report translates key insights from our 2025 AI survey, and share policy recommendations to help Canada make progress in establishing frameworks that give Canadians protections we deserve. We also remind the government to remain cautious about defaulting to disclosure alone as the answer to every gap we might see on these complex and extraordinarily fast-changing systems. More than transparency is needed today, and new approaches will be required as AI’s impact continues to evolve.

We point out that effective rights-driven regulation in practice will only exist with a strong and independent regulator charged with its enforcement. We challenge Canada’s current plan, especially on the Digital Safety and Data Protection Commission proposed in Bill C-36 –– Canada’s most recent attempt to reform private sector privacy law. We observe that the proposed super-regulator sits within a ministry whose principal interest is promoting Canada’s AI investment, putting enforcement in the hands of the same body shows a serious conflict of interests. At OpenMedia, we defend the role of independent Privacy Commissioner in enforcing privacy rights within AI-related matters.

Read our full submission to the public consultation here.

Summary of policy recommendations we proposed in the written submission

AI-generated content

Content of uncertain origin or unclear quality erodes public confidence in online spaces; AI slop makes it worse. Labelling AI-generated content is one of the most direct ways to mitigate the problem, yet inadequate alone. Therefore, we recommend our government:

Use both visible and invisible watermarks, including provenance metadata to make AI-generated content more identifiable, and to protect creative integrity by catching content that attempt to claim false sourcing or authenticity.
Establish a standardized framework with key identifiable icons and categorization to mitigate inaccurate labeling.
Support independent, fact-based, local journalism, and promote chain of custody efforts for non-AI content such as the Content Provenance and Authenticity (C2PA) voluntary standard to combat the proliferation of misinformation.
Promote AI literacy to empower Canadians to understand the risks and potentials of AI.
Introduce ex-ante regulation and implement a risk-based tiered regulatory system with precautionary legal measures that forbid AI from engaging in clearly harmful and manipulative behaviours.

AI interaction

The consultation explores how transparency around AI interactions could enable Canadians to make informed decisions. This may be partially possible, only if meaningful choices are available to exercise once disclosure happens. In Canada, we already see some of the world’s most consolidated telecommunications markets, and digital services are following the same path, becoming ever more concentrated in only a handful of Big Tech. As we see similar dynamics emerge in the AI industry, we stress that Canadians need more than just mandated disclosure for informed decision making measures, but mandated personal control. We recommend they:

Establish regulated ability for Canadians to consent or refuse consent, and to contest a decision made by an AI system and appeal to a human.
Require auditable, consumer-friendly design features built in AI systems to protect consumers when the costs of switching services are getting higher due to increasing monopolistic tendencies.
Reserve consumer-facing disclosure for moments where actions related to consequential decisions, data use, and points where a human alternative or appeal exists to avoid “disclosure fatigue.”

Information on AI systems

All members of our society hold their respective, legitimate reasons that justify when, why and what information they need on AI systems. But to our community, the foundation of AI system transparency should be clear ongoing public accountability. To meet different stakeholders, make sure AI’s economic benefits are responsibly realized, and our government puts the protection of Canadians’ rights and privacy at the forefront, we recommend our regulations:

Require plain language summaries of how AI systems are deployed accessible to everyone, as well as more technical public summaries for expert analysis, and full access to otherwise confidential material for accredited independent auditors.
Establish frameworks that scrutinize the full AI value chain with continuous oversight and well-documented records at every stage of development and deployment, since responsibilities are distributed across a chain of design, funding, and deployment decisions, and AI harms can no longer be traced easily to a single actor.
Mandate obligations and audits, and designate an independent regulatory body to monitor disclosure and accountability. We suggest four mechanisms, including designating senior accountable officers for compliance, mandating publication of a risk management framework, requiring access to full documented design and deployment decisions upon regulator request, and demanding high-risk systems be subject to independent third-party audit.

AI incidents

AI products are being widely deployed across sectors in Canada. Their adoption has tripled since 2024, and their associated risks are commensurately expanding. Yet, Canada still lacks a comprehensive legislative framework to govern AI, and conventional safeguards, such as human oversight, internal reporting, and the like, are being left to the individual practices of businesses or industries. To mitigate AI incidents that are growing increasingly hard to estimate and predict, we recommend:

Establish two complementary mechanisms: a systemic reporting regime, and developing government capacity to study emerging AI risks.
The mandatory reporting regime should assign responsibility according to who holds control at each stage of AI development and deployment, within a government-mandated format that adapted the concept of “real risk of significant harm” from PIPEDA.
Develop clear definitions that help stakeholders understand AI harms, and enable a direct path for civil society and individuals to report suspected violations and harms to the regulator.
Partner with domestic and international institutions for the study of AI risks, and align Canada’s reporting framework with emerging international standards in democratic peers that share Canadian values and principles.

AI agents

Agentic AI use is proliferating among Canadian organizations. A recent study by Oxford Economics and SAP revealed that 66% of Canadian businesses are already piloting agentic AI in their workflow, yet a staggering 97% are not fully prepared to govern the tools. The same study confirmed our concerns over the lack of human oversight, with 46% of respondents admitting that their organizations don’t have a human-in-the-loop process for autonomous workflows.

The lack of effective organizational oversight or intentionality in how it is using AI deepens the existing problematic status quo in which human end-users often have little to no control over how their data is being used as a condition of using a service. To uphold public trust and accountability of organizations, we recommend government:

Establish a robust reporting regime, and make organizations’ failure of compliance to reporting requirements consequential.
Mandate disclosure on the use of agentic AI when the implementation directly influences or impacts the financial, legal or life costs or opportunities of any individual Canadian.
Require organizations to perform independent systematic analysis for bias when AI agents are deployed to make structural decisions, such as education, hiring, tenancy or law enforcement, and guarantee a right to appeal to AI decisions.
Update consumer protection and contract law to clarify liability, require disclosure of the use of AI, and provide a right to human appeal.
Ban the use of agentic AI or other AI systems to contact or target voters in Canadian electoral law, with stiff penalties for violations to protect the integrity of our democracy.

Broader considerations

Getting AI transparency right will take several bites at the regulatory apple. We trust that with further studies and iteration, Canada will form a more complete governance blanket over time. We include the following points to wrap up our submission to reflect our community’s principles and philosophies, as well as some pragmatic considerations related to Canada’s recent initiatives on international cooperation.

AI transparency matters most when AI is involved in a consequential decision affecting a person’s life, and when a human might reasonably mistake an AI system for a real human. Therefore, obligations should scale with both risk and role, rather than applying uniformly across all organizations.
Canada should pursue interoperable standards with the EU and OECD to maintain coherence with jurisdictions that share Canadian values and principles, rather than the so-called laissez-faire, industry-driven regulatory posture adopted by the United States.
Our government should build in a mandatory two-year review period into AI regulation that includes public consultation, in order to respond to AI technology’s development growing at an exponential rate.
Ensure mandatory review and ongoing administration of regulatory frameworks are placed in the hands of a strongly independent regulator that is insulated from direct government interference, particularly from ministries that are simultaneously tasked with the promotion of AI investment.

For anyone who wants to learn more about our policy recommendations, you may find the full text of our written submission to the consultation here.

OpenMedia will continue advocating for right-respecting, people-first AI policy. Our work is 100% supported by people like you, so every dollar makes a huge difference for us. Consider supporting us today!