A new report by The Canadian Regulatory Technology Association, entitled noted AI continues to be a topical issue for regulatory actors, as the use of AI has increased and financial institutions (FIs) have successfully deployed AI in production or in proof-of-concept initiatives, the conversation has shifted from high-level principles towards how to effectively implement AI at scale. The report is based on expert viewpoints from regulatory technology space and focused on best practices to control and manage these risks.
The report noted, “over the last several years, FIs have developed and are developing principle-based guidelines describing factors to be considered in the deployment of machine learning (ML). These guidelines outline relevant questions and risks of ethics and discrimination with the objective of ensuring that these kinds of risks are being considered and adequately addressed”.
The main objective of this report is to introduce some common implementation challenges in terms of how they originate and manifest themselves within organizations, and then to delve into best practices on how to control implementation risks through sound governance practices and the use of tools.