The Bank of England, in its December 2023 Financial Stability Report and subsequent press conference yesterday, highlighted several risks associated with the continued adoption of artificial intelligence (AI) and machine learning (ML) in financial services. Specifically, the Bank expressed concern about:
- The potential for “black box” decision-making: The lack of transparency and explainability of AI and ML algorithms could make it difficult for regulators and supervisors to understand and assess the risks associated with these technologies.
- The concentration of power in the hands of a few large tech companies could lead to increased market concentration and reduced competition, which could be detrimental to financial stability.
- The potential for AI and ML to exacerbate existing inequalities: If biased algorithms are used to make decisions about credit, loans, and other financial products, this could lead to further disadvantages for certain groups of individuals.
The Bank of England also emphasized the need for:
- Clear and consistent regulatory frameworks: Regulatory frameworks should be developed to address the risks associated with AI and ML in financial services. These frameworks should be flexible enough to accommodate innovation and robust enough to protect financial stability.
- Strong governance and risk management practices: Financial institutions must have strong governance and risk management practices to manage the risks associated with AI and ML. This includes having clear lines of responsibility, conducting regular risk assessments, and implementing appropriate safeguards.
- Increased transparency and explainability: AI and ML algorithms should be designed in a transparent and explainable way. This will help to ensure that these technologies are used fairly and ethically.
- Collaboration between regulators, supervisors, and industry: Collaboration between regulators, supervisors, and the financial services industry is essential to ensure that AI and ML are used responsibly and in a way that promotes financial stability.
This is the first time that AI/ML has appeared in their Financial Stability Report, which highlights how this issue has elevated in the minds of financial market regulators. When regulators address issues they perceive to have systemwide impacts, they tend to take the most drastic action. It is incumbent on all of us operating in this space to continue educating key stakeholders to prevent overreaction to this emerging capability.