Addressing AI Operational Challenges
Date - February 1, 2022.
TORONTO, February 1, 2022 – The Canadian Regulatory Technology Association (CRTA) / Association canadienne de la technologie réglementaire (ACTR) today published a report entitled Moving Beyond Principles – Addressing AI Operational Challenges. It explores the current state of artificial intelligence and machine learning (AI/ML) adoption by examining how AI/ML is being used by financial institutions and the different governance approaches they are adopting to manage AI/ML risks. The paper delves into three challenges that organizations face when deploying AI/ML, bringing in perspectives of practitioners, academics, and technologists to share their experience, provide insights into current thinking, and offer practical guidance on how to deploy AI/ML effectively and safely.
Donna Bales, founder of the CRTA and lead author said: “The intent of this paper is to introduce some of the common implementation challenges and to identify best practices for controlling implementation risks through sound governance practices and tools. “She continued: “Recently, there has been a good convergence amongst regulated firms and regulatory bodies on AI risk themes and the value of principles to help manage these risks. However, the rate of which organizations are deploying AI/ML varies, and robust machine learning operations (MLOps) and testing tools are still quite nascent, resulting in variances between the frameworks and how they are assessed.” Adam Leon Smith, Chief Technology Officer at Dragonfly and one of the contributors to the paper commented: “Although the financial service industry has significant experience in managing model risk, the risks and controls needed for AI systems are fundamentally different. The regulatory drivers are also evolving from a focus on financial risk, to wider societal risks.”
The CRTA announced that, complementing publication of the white paper will be an invitation-only forum: Artificial Intelligence – A Road to Good Governance. The forum’s purpose is to bring together key stakeholders in the Canadian ecosystem so that they can collaboratively work together to identify best practices and collective approaches to strengthen AI governance, as well as address AI data challenges. Bales said: “Many financial service firms struggle to understand how AI should be managed and overseen by governance functions in key areas, such as operational resilience, data and model risk management. They recognize the need to strengthen culture and internal practices to support the safe use of AI.”
AI remains an important theme for CRTA members. As AI use evolves, the CRTA has a role to play both in improving understanding of best practices and MLops tools used to govern AI/ML and in enhancing commonality in the use of standards, definitions and certification methods.
The Canadian Regulatory Technology Association / Association canadienne de la technologie réglementaire
The Canadian Regulatory Technology Association is a not-for-profit organization focused on solving regulatory challenges through collaborative efforts among key RegTech stakeholders: regulated entities, technology vendors, regulatory bodies, government, and professional service providers. The goal is to facilitate dialogue, raise standards, and promote growth and innovation within the Canadian RegTech eco-system. The organization will endeavour to solve regulatory challenges through collaborative discussion and engagement in proof-of-concept initiatives.