The Influence of AI on the Canadian Financial Services Industry

The Department of Finance has launched a consultation to develop cohesive regulatory standards for Canada’s financial institutions, particularly regarding the integration of artificial intelligence (AI). These proposed regulations aim to strengthen consumer protections and address significant privacy and ethical challenges posed by AI. This article analyzes the Department’s proposals, explores AI’s potential in the financial sector, and discusses the need for structured oversight to balance innovation with responsible governance.

On August 12, 2024, the Department of Finance initiated a public consultation to gather feedback on proposed changes to Canada's regulations for financial institutions. This consultation represents the third phase of the Department of Finance's legislative review of the Bank Act, Insurance Companies Act, and Trust and Loan Companies Act, and related legislation and policies.¹  

The legislative proposals set out in the third phase fall under five themes; enhancing consumer protections, reducing geopolitical risks, introducing strong regulation, modernizing the financial sector framework and supporting a competitive market structure.²  

Amongst other key topics, the integration of artificial intelligence into the financial sector is explored. The consultation paper outlines the federal government’s initiatives to ensure the responsible integration of AI in the financial sector, including the creation of a federal strategy to maximize the benefits of AI while effectively managing risks.³

In recent years, Canadian banks and financial institutions, known for their expertise in AI research and implementation, have quickly increased their use of AI technologies to improve key business areas like lending, insurance underwriting, customer service, and risk management.⁴

Integrating AI into financial services could greatly improve the consumer experience by delivering personalized services, enhancing accessibility, and bolstering security by allowing for real-time fraud detection. Furthermore, increased automating can lead to faster service, thereby, increasing efficiency and sustainability.⁵  

While AI offers numerous advantages in financial services, its rapid innovation and implementation also present significant risks. These include potential data and privacy breaches due to the vast amounts of information stored in AI systems, challenges with regulatory compliance, and the risk of unfair outcomes stemming from biased data that AI systems might inadvertently learn from.⁶ 

At present, there are no enforceable AI regulations in the Canadian financial sector, and the current framework consists of a fragmented collection of non-binding white papers and reports. 

In 2017, the Office of the Superintendent of Financial Institutions (OSFI) released Guideline E-23: Enterprise-Wide Model Risk Management for Deposit-Taking Institutions, offering recommendations for AI oversight within financial institutions.⁷ 

More recently, the OSFI partnered with the Global Risk Institute (GRI) to form a community of AI thought leaders, bringing together experts from academia, regulatory bodies, banks, insurers, pension plans, fintechs, and research centers.⁸ This group, known as the Financial Industry Forum on Artificial Intelligence (FIFAI), pushed forward discussions on establishing effective safeguards and risk management practices for the use of AI in financial institutions.⁹  

In April 2023, the group published a report – A Canadian Perspective on Responsible AI. This report outlines five key principles designed to guide the responsible development of AI at Financial Institutions, known as the "EDGE Principles," which focus on Explainability, Data, Governance, and Ethics.¹⁰  

• Explainability should be factored in from the beginning of model design and, in certain cases, an explainable model may be chosen over a higher-performing opaque one, recognizing that modeling objectives can go beyond just performance.¹¹ 
• The integration of AI introduces challenges in handling larger volumes of data and managing new, diverse data sources. Thus, it is crucial for financial institutions to align their business and data strategies to ensure they are properly collecting, managing, and analyzing the right data to achieve their objectives. Effective data governance ensures that data is accurate, consistent, and complete, which is essential for the proper functioning of AI systems.¹²
• Governance has become increasingly important and should be holistic, covering all organizational levels with clear roles, responsibilities, and a defined risk appetite. It must also remain flexible as AI adoption matures. Specifically, AI model governance requires a multi-disciplinary approach and should be integrated into a risk-based culture, rather than treated as a routine exercise.¹³
• Ethics in AI is complex and subjective, as ethical standards change over time. Complying with the law doesn't always ensure fairness, and in some cases, bias is intentional, such as in pricing policies or risk stratification.¹⁴ AI training data can introduce bias, leading to unfair outcomes. To address this, financial institutions often follow a "fairness through unawareness" approach, avoiding the use or collection of certain personal attributes in decision-making.¹⁵ As seen lately, societal expectations for financial institutions to uphold strong ethical standards are steadily rising, and any harm—whether real or perceived—can lead to significant reputational risks and consequences. To mitigate this, organizations should ensure transparency externally and internally by disclosing how they maintain ethical standards in their AI models.

As previously emphasized, AI is widely acknowledged for its potential to offer significant benefits to financial institutions and their customers. Although the OSFI provides some guidance regarding safety and responsible deployment in this area, many stakeholders have stressed the need for stronger and more comprehensive federal leadership. As is evident, there is a pressing need for formal, centralized and binding regulation to ensure the safety and stability of the financial system. 

Furthermore, nationwide consistency, as outlined in Bill C-27, would provide valuable support in ensuring uniform regulations across the country. Bill C-27, known as the Digital Charter Implementation Act 2022, was introduced to strengthen Canada’s privacy laws and establish new regulations for the responsible development and deployment of AI.¹⁶

Currently tabled and ready for its third reading, this bill ensures that AI systems are deployed in a way that mitigate the risks of harm and bias, establishes an AI and Data Commissioner to oversee compliance and audits, and outlines clear criminal prohibitions and penalties regarding unlawful data use and AI deployment that causes harm or economic loss.¹⁷

Moreover, the Department of Finance, guided by an AI expert, is enhancing federal leadership on the responsible use of AI in the financial sector by engaging diverse stakeholders, evaluating potential risks, and formulating a federal strategy to maximize AI's advantages while minimizing associated risks.¹⁸ 

In summary, the growing use of AI in Canada's financial sector offers immense potential but also requires careful management of risks. Stronger federal leadership, formalized regulations, and nationwide consistency are essential for ensuring ethical, safe, and effective AI deployment. Continued collaboration between stakeholders and regulators will be key to shaping a secure and innovative future for the financial industry.

¹Consultation Paper: Proposals to Strengthen Canada’s Financial Sector - Canada.ca

²Consultation Paper: Proposals to Strengthen Canada’s Financial Sector - Canada.ca

³Consultation Paper: Proposals to Strengthen Canada’s Financial Sector - Canada.ca

⁴Consultation Paper: Proposals to Strengthen Canada’s Financial Sector - Canada.ca

⁵How AI is Shaping the Future of Financial Services in Canada (ncfacanada.org)

⁶How AI is Shaping the Future of Financial Services in Canada (ncfacanada.org)

⁷Enterprise-Wide Model Risk Management for Deposit-Taking Institutions - Guideline (2017) - Office of the Superintendent of Financial Institutions (osfi-bsif.gc.ca)

⁸Financial Industry Forum on Artificial Intelligence: A Canadian Perspective on Responsible AI - Office of the Superintendent of Financial Institutions (osfi-bsif.gc.ca)

⁹Financial Industry Forum on Artificial Intelligence: A Canadian Perspective on Responsible AI - Office of the Superintendent of Financial Institutions (osfi-bsif.gc.ca)

¹⁰Financial Industry Forum on Artificial Intelligence: A Canadian Perspective on Responsible AI - Office of the Superintendent of Financial Institutions (osfi-bsif.gc.ca)

¹¹Financial Industry Forum on Artificial Intelligence: A Canadian Perspective on Responsible AI - Office of the Superintendent of Financial Institutions (osfi-bsif.gc.ca)

¹²Financial Industry Forum on Artificial Intelligence: A Canadian Perspective on Responsible AI - Office of the Superintendent of Financial Institutions (osfi-bsif.gc.ca)

¹³Financial Industry Forum on Artificial Intelligence: A Canadian Perspective on Responsible AI - Office of the Superintendent of Financial Institutions (osfi-bsif.gc.ca)

¹⁴Financial Industry Forum on Artificial Intelligence: A Canadian Perspective on Responsible AI - Office of the Superintendent of Financial Institutions (osfi-bsif.gc.ca)

¹⁵Financial Industry Forum on Artificial Intelligence: A Canadian Perspective on Responsible AI - Office of the Superintendent of Financial Institutions (osfi-bsif.gc.ca)

¹⁶Bill C-27 summary: Digital Charter Implementation Act, 2022 (canada.ca)

¹⁷Bill C-27 summary: Digital Charter Implementation Act, 2022 (canada.ca)

¹⁸Consultation Paper: Proposals to Strengthen Canada’s Financial Sector - Canada.ca