On the use of artificial intelligence in financial regulations and the impact on financial stability

Jon Danielsson,Andreas Uthemann

2024-06-06

Abstract:Artificial intelligence (AI) can undermine financial stability because of malicious use, misinformation, misalignment, and the AI analytics market structure. The low frequency and uniqueness of financial crises, coupled with mutable and unclear objectives, frustrate machine learning. Even if the authorities prefer a conservative approach to AI adoption, it will likely become widely used by stealth, taking over increasingly high-level functions driven by significant cost efficiencies and superior performance. We propose six criteria for judging the suitability of AI.

General Economics,Risk Management

What problem does this paper attempt to address?

The paper aims to explore the application of Artificial Intelligence (AI) in financial regulation and its impact on financial stability, and proposes a series of standards for evaluating the applicability of AI. Specifically, the paper focuses on the following aspects:

1. **Malicious Use**: The financial system contains numerous resource-rich profit maximizers who may use AI to bypass regulations, alter the system for personal gain, and these actions are often difficult for other market participants or regulatory agencies to detect.

2. **Misinformation Propagation**: AI users may over-rely on it without understanding its limitations, especially in the process of macroprudential policy-making where data is scarce and goals are ambiguous. Data-driven algorithms may provide confident but inaccurate recommendations based on insufficient data.

3. **Goal Misalignment**: Ensuring that AI adheres to the goals expected by human operators is very challenging, particularly when faced with multiple and potentially conflicting objectives, such as maximizing profits, complying with the spirit and letter of the law, and ethical behavior. AI may prioritize goals in undesirable ways.

4. **Market Structure**: Companies that design and operate AI engines rely on three scarce resources: computing power, human capital, and data. This leads to increasing returns to scale, which hinders new entrants, increases homogeneity, and forms an oligopolistic market structure, making the system more susceptible to shocks.

The paper discusses how AI might exacerbate the vulnerabilities of the financial system through the analysis of the above four channels. It points out that although most regulatory agencies tend to adopt a slow and cautious approach to AI technology, this is not practically feasible. Additionally, the paper explores the issues faced by regulatory agencies in using AI themselves and proposes six standards for evaluating the use of AI to guide relevant regulatory activities.