The IIF has provided comment on the U.S. Federal Register Request for Information and Comment on Financial Institutions’ Use of Artificial Intelligence, including Machine Learning, published on March 31, 2021.
Our response provides data points and examples drawn from our previous IIF surveys on the use of machine learning in credit risk, anti-money laundering and financial crime prevention, and around the end-to-end governance of machine learning models.
We highlight that all models should be subject to an appropriate control framework that ensures that appropriate controls are in place commensurate with the risk of each specific use case, regardless of whether AI or ML techniques are used.
Our discussion provides input on explainability, and issues around data processing and usage, i.e. data use and accountability, and issues around siloed data and data quality. Additionally, we highlight the numerous steps that can be taken by FIs to tackle issues around unfair bias and ethical implications in ML.