Model Interpretability

A lender's model scores a loan application at a 79% risk of default, and the application is declined. In the United States that decision comes with a legal duty: Regulation B, which implements the Equal Credit Opportunity Act, requires the creditor to give the applicant the specific principal reasons for the adverse action (12 CFR 1002.9). In 2022 and 2023 the Consumer Financial Protection Bureau issued circulars applying that duty to complex algorithmic models; it withdrew both in May 2025, and the regulation itself did not change. In the European Union, the Court of Justice held in Dun & Bradstreet Austria (February 2025) that a person subject to an automated decision is owed an explanation of "the procedure and principles actually applied". The AI Act's Article 86 adds a right to an explanation of individual decisions made with high-risk systems, credit scoring among them, from 2 December 2027, the date set by the Digital Omnibus (Regulation (EU) 2026/1744). As of September 2026, "the model said so" is not an answer anyone can give.