Bank of America warns AI could widen a “credit divide” for consumers, study says

Bank of America’s view, as described in a market report published by TheStreet, is that artificial intelligence is not just changing financial markets and corporate operations. It could also create a “credit divide” that affects consumers, potentially determining who gets credit at reasonable terms and who faces tighter access. The post frames the issue as an uneven shift in lending. Rather than treating AI as a neutral upgrade to underwriting, it suggests that AI systems and the data they rely on can produce winners and losers, depending on what inputs are available and how models are deployed. The headline emphasis is on consumer impact, indicating that the bank sees second-order effects beyond traditional investor narratives about AI. While the report’s specific methodological details were not included in the material available for this write-up, the core concern is straightforward: credit is built on estimates of risk, and AI can change the speed, scale, and granularity of those estimates. If model outputs are calibrated using different quality or quantity of borrower information, the results can diverge across groups and circumstances, even when applications look similar on the surface. The “credit divide” framing also points to how lending systems can harden existing gaps. AI models can extrapolate from historical patterns, which means past behavior and past underwriting decisions may continue to influence who is deemed low risk and who is categorized as higher risk. If those classifications then drive pricing, limits, or approval likelihood, the system can reinforce the very differences it was trained to predict. For Bank of America, the issue sits at the intersection of growth, risk management, and regulatory scrutiny. Large banks use credit scoring and underwriting models to comply with capital and risk requirements and to make consistent decisions at scale. When AI is introduced into those processes, it can improve operational efficiency and decisioning accuracy, but it can also raise questions that regulators and the public are already focused on, including explainability, model governance, and bias. TheStreet’s summary places the emphasis on how AI may redistribute opportunities. That matters commercially because credit access influences consumer spending and household cash flow. In banking, even modest changes to approval rates or underwriting stringency can affect portfolio composition, charge-offs, and revenue mix. If some borrowers benefit from faster or more favorable decisions while others experience reduced access, the bank could see portfolio effects that differ from traditional expectations. Still, several details are not disclosed in the headline-level coverage available here. The market report does not provide, in the material reviewed for this article, the names of the underlying analysis, the timeframe, the specific AI techniques discussed, or how the bank measured “credit divide” outcomes. It also does not specify whether the concern is centered on Bank of America’s own model changes, the broader banking industry, or both. What to watch next is how this warning translates into concrete actions, such as enhancements to model testing, documentation of data provenance, and additional controls aimed at reducing unintended disparities. If Bank of America or other large lenders expand disclosures around model governance, or if regulators increase guidance on AI-driven credit decisioning, the practical meaning of a “credit divide” may become clearer to borrowers and investors alike.

keyFacts

whyItMatters

companies

tickers

sector

sourceTrail

confidence

needsReview