The OSFI Guideline E-23 model risk management will be fully effective by May 2027. For Canadian financial institutions (FIs), this means they need an enterprise-wide validation capacity to ensure every algorithm using data to generate output is a model that operates correctly and transparently.
What’s happening: The OSFI guideline aims to strengthen the risk management practices of FIs, specifically in relation to models used for credit, market, and operational risks. This is a significant change, as it requires institutions to adopt an enterprise-wide approach to model risk management, ensuring all models are properly developed, maintained, and monitored.
The demand for action: The exposure for Canadian FIs lies in their capability to accurately assess the risk associated with each AI-powered algorithm, as well as their ability to defend these decisions when faced with regulatory scrutiny. Furthermore, institutions must be prepared to accept liability for any errors or inaccuracies generated by these algorithms.
Action steps: To address this exposure, prudent institutions should take the following steps:
1. Conduct a comprehensive review of all existing models used for credit, market, and operational risks to identify potential vulnerabilities.
2. Develop an enterprise-wide model risk management framework to ensure consistency across all business lines and functions.
3. Implement robust monitoring and testing processes to continuously assess and mitigate any risks associated with these algorithms.
The difficulty: Institutions struggle to address these issues due to internal teams being too close to the decision-making process, and a lack of independent means to validate vendor claims.
Straven & Co’s role: Straven is an independent AI validation firm that helps organisations like yours by providing a plain verdict on proposed solutions. This involves testing against real operations, people, governance obligations, legal exposure, and regulatory expectations. By delivering a clear assessment, Straven ensures that organisations can make informed decisions about their AI investments.
Straven & Co examines AI decisions before they are acted on: stravenandco.com