AI Washing And The Imperative For Board Governance

Governance of artificial intelligence in lending has not kept pace with adoption, creating risks as institutions deploy complex models they cannot fully explain, according to a piece by lending professional Daniel Arhin. The article notes that while over 60% of financial institutions have implemented AI in at least one key function, traditional oversight frameworks designed for transparent models like scorecards are inadequate for opaque AI systems. It highlights regulatory responses such as the EU AI Act classifying credit scoring as high-risk and the UK's Financial Conduct Authority raising concerns about algorithmic bias. Real-world risks include the 2019 Apple Card scrutiny over alleged gender bias and model instability during the COVID-19 pandemic, which exposed AI systems' sensitivity to sudden data shifts. The piece concludes that innovation without effective oversight is not progress but risk, emphasizing that transparency is now a requirement in AI governance.