P2.T10.21.1. Machine learning (AI&ML) for risk management

Nicole Seaman

Director of FRM Operations
Staff member
Learning objectives: Explain the distinctions between the two broad categories of machine learning and describe the techniques used within each category. Analyze and discuss the application of AI and machine learning techniques in the following areas: Credit risk; Market risk; Operational risk; Regulatory compliance. Describe the role and potential benefits of AI and machine learning techniques in risk management. Identify and describe the limitations and challenges of using AI and machine learning techniques in risk management.


21.1.1. In regard to machine learning techniques, which of the following is a TRUE statement?

a. Regression implies a linear method
b. Principal component analysis (PCA) creates a curse of dimensionality by increasing the number of dimensions
c. A key feature of deep learning networks, and a source of risk, are the hidden layers located between the inputs and output
d. Support vector machines (SVM) and decision trees are unsupervised but they are unpopular because they are hard to explain

21.1.2. According to the authors (Aziz and Dowling),(†) artificial intelligence and machine learning (AI&ML) offer the powerful potential to reduce internal and external fraud. This fraud and fraud-related detection includes, but is hardly limited to, network analysis utilized to monitor employees; clustering techniques that establish behavior-based trader profiles; natural language processing (NLP); and techniques related to Know Your Customer (KYC). As an application, fraud detection and prevention belongs to which of the following risk areas?

a. Credit risk, in particular in the areas of consumer lending
b. Operational risk and/or regulatory compliance; aka, RegTech
c. Market risk, in particular the impact of trading on market price and other variables
d. FinTech risk and the risk of disruption by BigTech seeking to offer FinTech services

21.1.3. According to the authors (Aziz and Dowling),(†) there exist several challenges and practical issues that may limit the potential for artificial intelligence and machine learning (AI&ML) to realize its potential in its journey to transform the risk management function. In regard to these challenges, each of the following statements is TRUE except which is false or inaccurate?

a. Data silos: data is often collected in separate silos across departments and/or systems
b. Talent shortage: availability of skilled staff (i.e., skilled in AI&ML) is a top concern of firms
c. Opacity: many of the AI&ML techniques are effectively black boxes to a general audience
d. Impracticality: AI&ML is unlikely to help prevent unwarranted risks due to the inability to deliver real-time information

Answers here:

(†) Aziz, S. and M. Dowling (2019). “AI and Machine Learning for Risk Management”, GARP
Last edited by a moderator: