On February 5, 2024, the Conseil de l'innovation du Québec (CIQ) issued a report entitled Prêt pour l’IA : Répondre au défi du développement et du déploiement responsables de l’IA au Québec (available in French only). This report stems from a consultation process involving 250 experts and civil society organizations.
Recommendations for responsible AI in Québec
The report addresses various issues and opportunities associated with AI. It puts forward a series of proposals, including 12 main recommendations supported by 25 complementary recommendations aimed at ensuring the responsible development and use of this technology in Québec, grouped into five areas of focus: regulation, anticipation, education, support and positioning.
It also points out that AI should not be allowed to evolve without an appropriate legal framework. However, it notes that existing legal standards are often hard to apply to AI; the current legal framework has a number of gaps with respect to technological advances. It decries the negative consequences of the hands-off approach taken in recent years, including disinformation, discrimination and surveillance, all of which stand to grow worse over time.
Regulating AI with made-in-Québec legislation
To address the risks posed by AI, the report calls on Québec to adopt framework legislation to regulate AI development and implementation, drawing on the principles set out in the Montréal Declaration. It recommends basing Québec’s legislation on the severity of risks associated with AI systems, in keeping with the approach selected by the Canadian federal government and the European Union. It further urges that this legislation set standards for the use of AI systems in the private and public sectors and create an independent oversight body, which would also be tasked with recommending and drawing up related implementing regulations. In addition, the report underscores the importance of civic engagement in drafting the regulatory framework.
While the Artificial Intelligence and Data Act proposed by the federal government in Bill C-27 would only apply to private sector organizations engaging in international and interprovincial trade and commerce, the scope of Québec’s AI legislation should be sufficiently broad to encompass government agencies, private sector organizations and individuals, regardless of whether they are acting as developers, distributors or users of AI systems.
Here is a summary of the main recommendations (RP) in CIQ’s Prêt pour l’IA report.
REGULATE AI and ensure agile governance |
|
|
|
RP-1 |
Promptly begin work that will lead to the adoption of framework legislation to regulate the development and implementation of AI in society. |
RP-2 |
Delegate responsibility for implementing the framework legislation, as well as powers to recommend and draw up implementing regulations, to a body independent from the executive branch. |
ANTICIPATE changes ushered in by AI and encourage Québec institutions to adapt to them |
|
RP-3 |
Swiftly launch an initiative to review and update labour laws and social policies to ensure that they take AI’s rapid technological changes into account. |
EDUCATE all Quebecers so everyone can reap the benefits of AI while facilitating its responsible adoption |
|
RP-4 |
Adapt educational programs across all age groups (kindergarten through university) to foster students’ ability to use AI effectively and to hone their critical thinking skills. |
RP-5 |
Swiftly put in place a “train the trainer” program giving Québec’s intermediary organizations, college technology transfer centres (CCTTs), professional orders and other similar stakeholders access to the human resources they need to fast-track the upgrading of AI knowledge among workers and organizations provincewide. |
PROMOTE AI research and development and encourage Québec-based researchers and entrepreneurs to become responsible AI leaders |
|
RP-6 |
Ensure the excellence of Québec’s AI research network by making significant investments in promising research projects and knowledge mobilization activities across the AI spectrum. |
RP-7 |
Ramp up support aimed at helping companies gain distinction in the emerging market for responsible AI development while enabling them to develop and market tools specifically designed to regulate and increase the safety and reliability of AI, assess the robustness or quality of AI-driven products and services, certify privately-designed AI systems, etc. |
RP-8 |
Boost the computing power available to actors in the local AI ecosystem by supporting local entities developing Québec’s infrastructure network and by actively supporting the creation of an international cooperation organization dedicated to AI research – a CERN for AI – with an office in Québec. |
RP-9 |
Support the development of a high-quality Québec cultural database in French and various Indigenous languages. |
RP-10 |
Develop a Québec-wide strategy to derive value from digital data held by the private sector, researchers, public authorities and civil society in order to facilitate data sharing between these stakeholders. |
POSITION Québec as an AI leader and role model |
|
RP-11 |
Immediately launch the “IA pour le Québec” [AI for Québec] program by providing government departments, public bodies, state-owned enterprises and municipalities with a significant budgetary envelope for the responsible planning and implementation of highly strategic AI projects. |
MOVE FORWARD with all these efforts |
|
RP-12 |
Immediately create an interim AI governance steering committee to move forward with efforts to appropriately regulate AI and to foster the conditions for the successful integration of AI within Québec society. |
In light of the major issues raised by the use of AI, the rigorous approach taken by CIQ to outline a legal framework for AI in Québec is especially praiseworthy. The tabling of a new bill, however, will create interoperability challenges with other draft legislation currently under study in various jurisdictions. Companies seeking to reap the potential benefits of AI would do well to keep a close eye on these legislative developments.
The authors acknowledge the contribution of Cléa Jullien, articling student, to this publication.