The partnership between OSFI and GRI created the Financial Industry Forum on Artificial Intelligence (FIFAI) which gathered Canada’s financial services experts from industry, government and academia. FIFAI culminated in the release of the EDGE principles (Explainability, Data, Governance and Ethics) for the responsible application and regulation of AI in the financial services.
Mizdrak highlighted AI’s potential to enhance business productivity and efficiency through data processing, but cautioned about inherent risks such as bias, content manipulation, and those related to the complexity of algorithms. She emphasized that the principles are based on risk management best practices and meant to assist institutions in evaluating AI systems by offering a consistent approach for assessing risks and benefits. She also stressed the need for regulatory efforts to keep up with the rapid pace of AI development through an iterative process.

Rubtsov went on to summarize the EDGE principles. The Explainability of a given AI system or function is critical to building and maintaining trust between a financial institution and its customers, but must be balanced with the need for confidentiality. The Data principle emphasizes data quality, consent, and privacy, with employee training reducing associated risks. Governance should be adaptable and reflect use cases, drawing on multi-disciplinary expertise. Ethics depends on business culture and societal values, with legal obligations requiring careful consideration.

The EDGE principles can be an important tool for financial institutions grappling with new opportunities to employ AI. Mizdrak concluded by affirming the regulator’s commitment to ongoing collaboration with various industry stakeholders, domestic and international peers, academia, and the private sector. The goal is to foster connections and develop additional AI guidance in the near future.