This report explores the five key areas of concern raised over hundreds of research interviews with financial services leaders and regulators conducted over the past two years. Specifically: Explainability, how does business context shape what we need to know about our AI, and how AI systems make their decisions? Systemic risk: Could algorithms destabilize the financial system by introducing new dynamics to how markets and their component actors (i.e., humans and machines) interact? Bias and fairness: How can institutions work to ensure their systems do not discriminate against a specific group? Fiduciary duty: Can AI systems plausibly fulfill the fiduciary responsibilities of advisors to always act in the best interest of their clients? Algorithmic collusion: Can AI systems autonomously learn to engage in anti-competitive behaviour, and how can this risk be mitigated?
The report outlines responses to these questions, exploring the root causes that drive each concern, and how they can be addressed through new governance and oversight frameworks. Ultimately, the report seeks to map out how the financial system can reap the benefits of AI while effectively navigating the risks it introduces, leading to improved outcomes for customers, regulators, institutions, and societies at large.