Responsible Use of AI in Financial Services: Balancing Risks and Opportunities

Artificial intelligence (AI) creates opportunities in financial services for greater efficiency, improved decision-making and enhanced customer service. However, as financial institutions (FIs) integrate AI models into their operations, they must navigate a number of risks inherent in the technology in order to maintain customer trust, and safeguard the security and stability of the financial system.

Large language models (LLMs) and AI agents are two applications of AI that promise to transform the financial sector. While LLMs are already in broad use for generating text-based outputs, AI agents expand on these capabilities by autonomously executing multi-step tasks with minimal human intervention. Both applications, but particularly AI agents, raise concerns about data privacy, security, and fairness that will need to be addressed.

In 2022 GRI and OSFI organized the Financial Industry Forum on Artificial Intelligence, where the EDGE principles—Explainability, Data, Governance, and Ethics—were developed to guide responsible AI implementation in the financial sector. This report expands on those core principles by integrating recent advancements in AI and addressing broader considerations such as AI literacy and environmental impact.

Ultimately regulations will play a vital role in setting the guardrails that ensure AI is developed and deployed responsibly, and this report also reviews several different regulatory approaches being considered. Some regulators favour a domain-specific model that leverages existing frameworks; others propose a framework approach to integrate AI principles across multiple sectors; still others pursue a whole-of-economy strategy with a single, comprehensive law. The report presents advantages and disadvantages of each approach.