The process for stakeholder involvement leading to the adoption of Ontario’s Bill 194 highlights significant shortcomings in meaningful public participation, particularly in the context of governing transformative technologies like AI. With the proroguing of parliament in January 2025, leading to the demise of the federal bill C-27 and the Artificial Intelligence and Data Act (AIDA), Bill 194 is Canada’s only significant AI legislation. While the government sought input through consultations and committee hearings, the process ultimately fell short of fostering genuine collaboration to shape the legislation. This failure not only undermined the democratic legitimacy of the bill but also squandered an opportunity to build public trust in AI governance, a critical issue as governments increasingly adopt AI systems with far-reaching social implications. With the bill being passed and the upheaval of the AI ecosystem globally, it is beyond time that the government of Ontario create meaningful participation opportunities for stakeholders that demonstrably impact AI adoption in the public service.
Bill 194 was tabled on May 13, 2024, and the government simultaneously released a consultation document on the Ontario Regulatory Registry seeking stakeholder feedback until June 11, 2024. Amongst others, the Information and Privacy Commissioner of Ontario,[1] Law Commission of Ontario,[2] Teresa Scassa,[3] and Open North, together with Brenda McPhail, provided detailed feedback. These responses almost uniformly agreed about both the content and the considerable scale of changes needed to make Bill 194 a leading piece of AI regulation. After this feedback period, and following its Second Reading in the Legislative Assembly, the bill was referred to a standing committee. This was a second opportunity for stakeholders to provide written or in-person comments. Unfortunately, the final form of Bill 194 does not reflect the feedback from either consultation opportunity and was largely passed unchanged from its original form. The benefits of public participation in government decision-making has long been recognized in scholarly and practitioner literature. Research is clear that with the appropriate sociopolitical and administrative structures, decisions made through participatory processes are viewed as more legitimate with increased acceptance of the resultant policy and less need for redesign. However, the consultation process enacted by the Ontario provincial government did not provide a meaningful opportunity to participate in legislative development, let alone explore issues of power, trust, and participation in the regulation of digital technology.
To understand these shortcomings, we can examine the process through established participation frameworks, such as Arnstein’s Ladder of Participation[4] and the International Association for Public Participation (IAP2),[5] as well as emerging literature on public engagement in AI governance. These tools reveal that Ontario’s approach to public participation was largely tokenistic. According to Arnstein’s Ladder, which categorizes levels of public involvement in decision-making, the process for Bill 194 operated at lower rungs, such as consultation and placation. While stakeholders were invited to provide feedback—primarily through the Ontario Regulatory Registry and committee hearings—there was no obligation for the government to act on their input. For instance, recommendations from the Law Commission of Ontario, which emphasized transparency and accountability measures, were largely ignored. This reflects a “window-dressing” approach where public input is solicited but not meaningfully integrated into decision-making.
Similarly, when analyzed through IAP2’s Public Participation Spectrum, Ontario’s process failed to meet even basic standards for effective engagement. Even the low end of the spectrum, “Consult,” requires the government to provide feedback on how public input influenced the decision. In this case, there is no indication that public feedback was seriously considered or incorporated into the final legislation. Participants were left in the dark about whether their concerns, such as calls for independent oversight of AI systems, were acknowledged or dismissed. This lack of transparency eroded trust and reinforced perceptions that public participation was merely a formality.
The need for meaningful public engagement is particularly acute when it comes to AI governance. As highlighted by leading organizations like the Partnership on AI[6] and the Ontario Human Rights Commission,[7] public participation is essential to ensure that AI systems are designed and deployed in ways that align with societal values and protect human rights. Research consistently shows that participatory processes help identify risks, mitigate harms, and build public understanding of complex technologies. Scholars such as Cohen and Suzor argue that democratic control over AI requires multiple channels for contestability, such as opportunities for third parties to access information, challenge decisions, and participate in audits.[8] Yet the development process for Ontario’s Bill 194 did not include meaningful participation, and furthe,r its content does not support stakeholders playing any meaningful role in ongoing oversight either. This has the effect of centralizing power within government institutions without sufficient checks or balances.
The consequences of this tokenistic approach to the development of Bill 194 extend beyond the bill itself. Public trust in government institutions is already fragile; recent surveys indicate that Canadians are losing confidence in both federal and provincial governments’ ability to address key issues effectively.[9] While Canadians have historically reported higher trust levels compared to global averages, this trust is eroding amid perceptions of opaque policymaking on emerging technologies like generative AI.[10] Given the unknowns surrounding AI’s societal impacts and its capture by large technology companies, the need for transparent and inclusive governance has never been more urgent.
By failing to engage meaningfully with the public on Bill 194, Ontario missed a critical opportunity to build trust and foster buy-in for its integration of AI into governance. The absence of iterative dialogue or collaborative problem-solving mechanisms left participants feeling excluded from decisions that could profoundly affect their lives. This lack of democratic legitimacy is particularly troubling given the high stakes of AI adoption, where risks to civil and human rights are significant. Ultimately, while Ontario’s Bill 194 represents an important step toward regulating AI use in the public sector, its flawed participatory process undermines its potential effectiveness. Meaningful public engagement would have required not only soliciting feedback but also following up with participants to interrogate their perspectives further and informing them about how their input shaped, or failed to shape, the final legislation. Without such efforts, public participation remains superficial, contributing to a decline in institutional trust at a time when governments must demonstrate their capacity to govern transformative technologies responsibly.
As both Arnstein’s Ladder and IAP2 emphasize, participation must go beyond consultation or placation to achieve genuine collaboration or empowerment. The emerging literature on AI governance reinforces this point: maintaining democratic control over high-risk technologies requires robust participatory frameworks that prioritize transparency, contestability, and inclusivity. Ontario’s experience with Bill 194 serves as a cautionary tale about what happens when these principles are neglected and a reminder that meaningful engagement is not just a procedural formality but a cornerstone of effective governance in the age of AI.
The failure to do so is all the more inexplicable given not just the plethora of research and projects on AI public engagement, but existing work from the Ontario government on potential AI policy that contradicts Bill 194 and its process. The detailed public engagement processes the province undertook as part of the development of the Trustworthy AI Framework in 2022 found that the public “wants to be involved in decisions related to how AI is developed, when it’s used, how it is trained” and also to “[e]ngage with sector leaders and civil society to develop a standard for ‘trustworthy AI.’”[11]
Indeed, the 2022 engagement as a whole was excellent work and contained extensive and useful findings (like responsible AI principles, principles that were notably missing from Bill 194); it is unclear why it was not built upon. Doing so would have advanced the process up the engagement ladder towards proper collaboration. Ontario had a lot to gain by positioning itself amongst the leaders in responsible AI governance alongside the likes of Amsterdam and Barcelona. Their processes have engendered a rich and productive civic technology community and brought equitable technology benefits and innovation to their societies.
The adoption of Ontario’s Bill 194 represents a missed opportunity to set a global standard for responsible AI governance through meaningful public participation. While the bill addresses critical issues related to cybersecurity, AI in the public sector, and digital technology’s impact on society, its development process failed to reflect the principles of transparency, inclusivity, and accountability that are essential in governing high-risk technologies. By disregarding detailed stakeholder feedback and neglecting to close the feedback loop, the Ontario government undermined trust in its institutions and weakened the democratic legitimacy of the legislation. The failure to build on previous engagement efforts, such as the Trustworthy AI Framework consultations in 2022, further highlights a lack of continuity and vision in Ontario’s approach to AI policy.
However, with much of Bill 194’s substance left to be determined through regulations, there remains an opportunity to course-correct by embedding meaningful public participation into the regulatory process. While the bill does not contain public engagement requirements, its preamble does, stating that “[t]hese regulations would be developed through consultations with government and public sector stakeholders, Indigenous partners, and AI experts.” The Ontario government must seize this opportunity to rebuild public trust and demonstrate leadership in AI governance by committing to a robust and inclusive regulatory consultation process. Ontario should endeavour to position itself as a model for responsible AI governance, one that not only protects human rights but also empowers the public to shape the future use of transformative technologies.
Sources:
[1] https://www.ipc.on.ca/en/media-centre/blog/bill-194-ontarios-missed-opportunity-lead-ai
[2] https://www.lco-cdo.org/wp-content/uploads/2024/06/LCO-Submission-to-Government-of-Ontario-Bill-194-Consultations-June-2024.pdf
[3] https://www.teresascassa.ca/index.php?Itemid=80&id=383%3Asubmission-to-consultation-on-ontarios-bill-194-strengthening-cyber-security-and-building-trust-in-the-public-sector-act-2024&option=com_k2&view=item&
[4] https://citizenshandbook.org/arnsteinsladder.html
[5] https://cdn.ymaws.com/www.iap2.org/resource/resmgr/pillars/iap2_spectrum_2024.pdf
[6] https://partnershiponai.org/stakeholder-engagement-for-responsible-ai-introducing-pais-guidelines-for-participatory-and-inclusive-ai/
[7] https://www3.ohrc.on.ca/en/bill-194-strengthening-cyber-security-and-building-trust-public-sector-act-2024
[8] https://policyreview.info/articles/analysis/contesting-public-interest-ai-governance
[9] https://www.environicsinstitute.org/projects/project-details/trusting-federal-and-provincial-government-decision-making-on-key-issues
[10] https://www.oecd.org/en/publications/2024/06/oecd-survey-on-drivers-of-trust-in-public-institutions-2024-results-country-notes_33192204/canada_1769aff6.html
[11] https://www.ontario.ca/page/ontarios-trustworthy-artificial-intelligence-ai-framework-consultations