A peer-reviewed qualitative study based on 35 semi-structured interviews across engineers, lawyers, regulators, and third-party vendors finds that US financial institutions universally maintain dedicated fair-lending teams and formal bias-testing procedures, but the specific methods used vary widely across firms. Regulatory supervision via examination, rather than complaint-driven enforcement or voluntary best-practice adoption, is identified as the decisive structural mechanism that produces compliance; soft requirements without supervisory authority are found to generate performative rather than substantive fairness work. For financial institutions building or procuring AI-based lending or credit models, this documents the compliance floor and the enforcement mechanism that defines it, and signals that institutions outside covered-entity definitions face structurally different and lower fairness-testing obligations.
113 other tracked developments touching the same claims.