Background on Canada's Digital Safety Act and AI Chatbot Provisions

The Canadian government recently introduced Bill C-34, known as the Digital Safety Act or Safe Social Media Act, which extends regulatory oversight to artificial intelligence chatbots alongside social media platforms. This legislation emerged in response to growing concerns over youth mental health and safety, particularly following high-profile incidents involving harmful interactions with conversational AI tools. The bill imposes duties on operators of public-facing conversational chatbots to mitigate risks of harmful content or behaviour, including requirements for transparency plans, crisis intervention protocols, and measures to prevent manipulative engagement techniques that could encourage social withdrawal or emotional dependency.

Under the proposed framework, chatbot services must address issues such as posing as licensed professionals, using techniques that foster unhealthy attachments, or providing advice that could lead to self-harm. A new Digital Safety Commission would oversee compliance, with powers to enforce standards and potentially exempt platforms that demonstrate adequate safeguards. This approach builds on earlier efforts like the stalled Artificial Intelligence and Data Act (AIDA) from Bill C-27, which failed to pass in 2025, leaving Canada with a patchwork of privacy laws, provincial rules, and voluntary guidelines.

Academic Experts Highlight Shortcomings in the Proposed Framework

Leading Canadian scholars have expressed reservations about the bill's scope and implementation. Evan Light, an associate professor at the University of Toronto specializing in technology and privacy, described the legislation as underdeveloped, noting that restrictions could be easily bypassed through tools like virtual private networks. He emphasized that the bill's preview suggests limited practical utility without more robust enforcement mechanisms.

Florian Martin-Bariteau, director of the Centre for Law, Technology and Society at the University of Ottawa, warned that young users might shift to unregulated smaller platforms, undermining the law's protective intent. Michael Geist, Canada Research Chair in Internet and E-commerce Law at the University of Ottawa, critiqued the "trust us" model, highlighting uncertainties around age verification requirements and the broad definition of regulated chatbot services that could encompass productivity tools alongside companion-style applications.

These critiques underscore a broader concern that the legislation risks repeating patterns from previous online harms bills by prioritizing rapid action over precise, evidence-based rules tailored to the unique one-to-one nature of chatbot interactions.

Implications for Canadian Higher Education Institutions

Universities across Canada are already navigating generative AI tools in teaching, research, and student support services. The new regulations could influence how institutions develop their own AI policies, particularly around academic integrity and mental health resources. Many universities, including those in the U15 group, have introduced guidelines for AI use in coursework, but the federal framework may require alignment with national safety standards for any chatbot-like tools used in advising or tutoring platforms.

Research at institutions like McGill University and Simon Fraser University has explored youth perspectives on AI chatbots through deliberative processes, identifying risks such as relational dependence and cognitive impacts. These findings suggest that higher education leaders must consider how regulatory duties might affect campus innovation hubs and AI ethics programs.

Perspectives from University Administrators and Faculty

Administrators at Canadian colleges and universities have noted the potential for increased compliance costs and the need for cross-institutional collaboration. Faculty members involved in AI literacy initiatives highlight the importance of preserving academic freedom while addressing safety concerns. Some institutions are already piloting internal AI governance frameworks that emphasize transparency and risk assessment, which could serve as models for broader adoption.

Concerns include how the bill's definition of "harmful behaviour" might intersect with educational uses of AI for personalized learning or mental health chatbots provided through student services.

Photo by Mohamed Nohassi on Unsplash

Student Experiences and Equity Considerations

Students at Canadian universities have voiced mixed reactions, appreciating efforts to protect vulnerable users while fearing overreach that could limit access to beneficial AI tools. Equity issues arise particularly for international students and those in remote programs who rely on AI for language support or academic assistance. The legislation's focus on major platforms may leave gaps for smaller educational technology providers commonly used in higher education settings.

Challenges in Implementation and Enforcement

Key challenges include defining the boundary between regulated companion chatbots and unregulated productivity tools, ensuring privacy protections remain intact, and building institutional capacity for audits and oversight. Experts note that without independent data access and algorithmic review powers, enforcement may prove difficult. The bill's reliance on a new commission raises questions about expertise and resourcing in a rapidly evolving technological landscape.

Comparative International Approaches

Canada's efforts can be compared to developments in the European Union, United Kingdom, and Australia, where targeted codes of practice and safety duties for chatbots have been advanced. These models emphasize pre-deployment assessments and ongoing monitoring, elements some Canadian academics argue are missing from the current proposal. Aligning with international standards could enhance Canada's competitiveness in responsible AI development.

Opportunities for Higher Education Leadership

Canadian universities are well-positioned to contribute to policy refinement through research partnerships and pilot programs. Initiatives at places like the University of Ottawa's Centre for Law, Technology and Society and Mila in Montreal demonstrate expertise that could inform more nuanced regulations. Collaborative efforts across provinces could help harmonize responses to the federal framework.

Photo by Emiliano Vittoriosi on Unsplash

Future Outlook and Recommendations

As the bill progresses through Parliament, ongoing consultations with academic stakeholders will be essential. Recommendations include strengthening transparency requirements, clarifying definitions to avoid unintended restrictions on educational tools, and investing in AI literacy programs within higher education. A balanced approach could foster innovation while safeguarding students and researchers.

Ultimately, the success of these regulations will depend on their ability to adapt to technological change without stifling the very advancements that position Canadian institutions as global leaders in AI ethics and application.