Download PDF:
March 2025 Advocacy Newsletter
Summary:
We’re just wrapping up our response to the Canadian Securities Administrators’ consultation on the use of artificial intelligence in capital markets. Until now, securities regulators’ responses to AI have come largely piecemeal, with one-off guidance and rules dealing with particular applications of this technology—some of which is over a decade old. This consultation has given us a chance to take a step back and see how well these responses fit together.
One area where we could use a lot more coherence is when it comes to human oversight of AI systems. When a firm uses AI to recommend investments, regulators demand that these recommendations be reviewed line-by-line by proficient individuals. But discount brokers that employ gamification and other digital engagement tactics that could steer their clients into some investments over others are allowed to operate without this friction.
It seems wasteful to require human review of all outputs from an AI system, no matter how low-stakes, but unduly risky to take humans entirely out of the loop. There’s an obvious middle ground, though—when an AI system seems likely to influence retail investor behaviour or affect market stability, proficient individuals ought to be involved in their design and oversight, and should periodically review the system’s outputs to ensure it’s working as intended.
As AI evolves along with its applications in investment management and financial services, regulators will need to respond quickly in a way that reflects our developing understanding of its risks and opportunities. Hopefully, the information collected through this consultation will leave them better placed to meet this challenge.