Canada Continues Consultation on AI Transparency: What Nonprofits Need to Know
By CENA | August 2026
Artificial intelligence is becoming increasingly embedded in the way Canadians work, communicate, access services and make decisions. As adoption accelerates, the Government of Canada is asking a fundamental question: How can Canadians know when they are interacting with AI, understand what AI systems can and cannot do, and hold organizations accountable when things go wrong?
On July 23, 2026, the Government of Canada launched a public consultation on strengthening transparency around artificial intelligence. The consultation is open until September 23, 2026, and is part of Canada's broader AI for All strategy, launched in June.
For Canada's nonprofit sector, the consultation is particularly relevant. Nonprofits are increasingly using AI for communications, research, administration, fundraising, service delivery and data analysis. Greater transparency requirements could influence not only how organizations use AI internally, but also how they communicate with clients, donors, volunteers and communities about that use.
Why AI transparency matters
AI systems are no longer limited to experimental technology departments. They are increasingly present in everyday products and services, while generative AI tools can create text, images, audio and video that may be difficult to distinguish from human-created content.
At the same time, newer AI agents are moving beyond generating information. They can potentially perform tasks, interact with digital systems and take actions on behalf of users.
The Government argues that greater transparency can help Canadians make informed decisions, strengthen trust and accountability, and provide organizations and policymakers with information needed to understand the impact of AI.
The consultation therefore focuses on five areas.
Five areas at the centre of the consultation
1. Identifying AI-generated content
One of the most visible questions is whether people should be able to tell when content has been generated or altered by AI.
The Government is considering approaches such as watermarks, disclaimers and provenance metadata that could help people identify AI-generated or manipulated content. It is also asking when this information matters most—for example, with images, video or audio.
For nonprofits, this could become increasingly important as organizations use AI to produce newsletters, social media graphics, videos, fundraising materials and educational resources.
Transparency does not necessarily mean avoiding AI-generated content. Instead, it may mean being clear about when and how AI was involved.
2. Knowing when you are interacting with AI
The consultation also examines whether people should be informed when they are interacting with an AI system.
This could have direct implications for nonprofits using AI-powered chatbots, virtual assistants or automated support tools.
Imagine a community member asking a nonprofit about housing resources, employment support or social services. Knowing whether the first interaction is with a person or an AI system can affect expectations about accuracy, privacy, escalation and accountability.
Interestingly, the federal government is already applying this principle to AI applications on Canada.ca. Government guidance states that AI help applications should be clearly labelled as AI and should provide information about privacy, potential mistakes, limitations and how user data will be handled.
This provides a useful signal for organizations considering their own AI deployments: transparency should be designed into the user experience, rather than added after something goes wrong.
3. Making AI systems easier to understand
Transparency is not only about saying, "This is AI."
The Government is also considering whether people should have better access to consistent and understandable information about AI systems—including their development, capabilities and limitations.
For nonprofit leaders, this raises an important practical question:
Before adopting an AI tool, do we understand what it can do, what it cannot do, what data it uses, and where its limitations could affect our clients or organization?
AI literacy therefore becomes part of organizational governance.
A team does not need to understand the technical architecture of a large language model. But decision-makers should understand enough about an AI system to assess whether it is appropriate for a particular task.
4. Tracking serious AI incidents
The consultation also proposes improving the ability to track serious incidents involving AI systems.
This is significant because responsible AI cannot simply focus on successful use cases. Organizations also need mechanisms for recognizing, documenting and learning from failures.
For nonprofits, this could mean thinking more systematically about questions such as:
What happens when an AI tool provides incorrect information?
Who is responsible for reviewing a problematic output?
When should an AI-generated decision or recommendation be escalated to a human?
How should significant incidents be documented?
What should clients or affected communities be told?
Incident reporting can turn individual mistakes into organizational learning—provided organizations have processes in place to capture and review them.
5. Greater accountability for AI agents
Perhaps the most forward-looking part of the consultation concerns AI agents.
Unlike conventional generative AI tools that primarily respond to prompts, agentic systems can carry out sequences of tasks and interact with other digital systems.
That creates new questions about traceability and accountability.
If an AI agent sends an email, updates a database, publishes information or performs another action, organizations need to know what the agent did, why it did it, what permissions it had and who was responsible for its operation.
The Government's own guidance on agentic AI emphasizes principles such as bounded autonomy, clear accountability, human oversight and the ability to stop or disable an agent when necessary.
For nonprofits experimenting with increasingly autonomous AI workflows, these principles are particularly relevant.
The more an AI system can do, the more important it becomes to know who is accountable for what it does.
What does this mean for Canadian nonprofits?
The consultation is still open, and the Government has not predetermined what specific measures will ultimately be adopted. It is seeking feedback on whether additional action is necessary and, if so, whether voluntary approaches, industry standards, legislation or a combination of measures would be most appropriate.
Nevertheless, nonprofits should not wait for regulation before thinking about transparency.
A practical starting point is to establish a simple internal AI transparency framework.
Know your AI.
Identify which AI tools your organization uses and what functions they perform.
Tell people when AI is involved.
Where appropriate, make it clear when clients, partners, donors or community members are interacting with an AI system.
Know the limitations.
Staff should understand that AI outputs can contain errors and should not automatically be treated as authoritative.
Keep humans accountable.
AI can assist with work, but organizations should clearly identify who remains responsible for important decisions and communications.
Document significant problems.
Create a process for reporting, reviewing and learning from serious AI-related incidents.
Be especially careful with AI agents.
As systems gain the ability to take actions rather than simply generate content, organizations should consider permissions, human checkpoints, traceability and the ability to intervene.
Transparency could become a competitive advantage
AI governance is sometimes viewed as a compliance exercise. But for nonprofits, transparency can also be a trust-building strategy.
Community organizations often work with people who are sharing sensitive information or relying on organizations during vulnerable moments. Trust is therefore not an optional feature of nonprofit AI adoption.
Being transparent about AI use can help organizations demonstrate that they are not simply adopting technology because it is available. They are adopting it deliberately—with clear boundaries and with the interests of their communities in mind.
This is especially important as AI becomes more capable.
The question is gradually moving from:
"Can AI do this?"
to:
"Should AI do this—and under what conditions?"
That shift represents a more mature stage of AI adoption.
CENA perspective: Transparency must become part of AI literacy
For the nonprofit sector, Canada's consultation is an important reminder that AI adoption and AI literacy must advance together.
Organizations need more than access to AI tools. They need the knowledge and governance practices required to use those tools responsibly.
Transparency should therefore not be treated as something that belongs only to technology companies or governments. It should become part of everyday nonprofit practice—from procurement and staff training to client communication and program delivery.
Canada's consultation remains open until September 23, 2026. The Government is inviting Canadians, businesses, researchers, civil society organizations, Indigenous groups and other stakeholders to share their views. Following the consultation, the Government plans to review submissions and publish a What We Heard report.
For nonprofits, this is an opportunity not only to respond to emerging policy, but to help shape what responsible AI adoption should look like across Canada's social sector.
The future of nonprofit AI will not be defined simply by how much technology organizations adopt. It will also be defined by how transparently, responsibly and thoughtfully they use it.
Participate in the Government of Canada’s AI transparency consultation.