Financial institutions that leverage artificial intelligence (AI) and machine learning (ML) are not in favor of new regulations determining how the techniques can be used.

The Federal Reserve and the Office of the Comptroller of the Currency (OCC) together with the Federal Deposit Insurance Corporation (FDIC), Consumer Financial Protection Bureau and the National Credit Union Administration (NCUA) have received replies to their April call for input from financial institutions, trade associations, consumer groups and other stakeholders.

Banks, tech companies, and bankers’ associations had previously filed comments with the OCC in response to an advance notice of proposed rulemaking (ANPR) the agency put out when it was “reviewing its regulations on bank digital activities,” in July 2020.

In its response, the Bank Policy Institute (BPI), a financial services industry group, said it “believes that new regulations are not necessary, and that the Agencies should apply a flexible, principled, and risk-based approach for the risk assessment, implementation and oversight of AI.” The research and advocacy group counts financial institutions like Bank of America, Capital One, Barclays, BNY Mellon and BNP Paribas among its members.

“Consumers are best served by regulatory approaches that are not static or rigid but are sufficiently flexible and adaptable to the emergence of new technologies,” BPI noted in its statement, adding that the agencies should avoid “creating or applying new regulatory expectations that may hinder progress” in implementing AI use cases in the financial services industry.

AI and ML should be treated like other technology and, rather than focusing on these techniques specifically, regulators should base their approach on whether they introduce any new risk into the system, the BPI said. Emphasizing this approach for explainability requirements surrounding AI, the industry group noted that “banks may hold AI models that are customer-facing, such as those used for consumer lending, to a higher standard of external explainability than AI models that are used for internal operational processes, such as processing documents.”

Echoing the BPI’s sentiments, Mastercard noted in its own response that, while “further regulation specific to AI isn’t necessary,” regulatory agencies should partner with Standard Setting Organizations (SSO) to formalize standards for third parties that provide banks with models that rely on AI, machine learning and big data processing.

Meanwhile, with regard to fair lending concerns in AI-driven credit underwriting raised by regulators, the nonprofit Mitre Corporation noted in its response that “there are clear demonstrations that AI can result in discrimination against protected classes,” and that developers of consequential ML systems need to emphasize fairness risks across the model’s lifecycle. Mitre manages federally funded research and development centers supporting various U.S. government agencies.

While legal and policy approaches to mitigating protected class discrimination designed to counter human bias are poorly suited for ML models, Mitre noted, “regulators may be able to promote improved, industry-wide AI maturity by providing supervised financial institutions with incentives such as increased regulatory cooperation on unintended outcomes involving AI-enabled applications.”

Although most respondents to the federal regulators’ request for public input seemed to urge agencies to not frame prescriptive rules governing AI in financial services, underwriting software provider Zest AI urged the agencies to clarify that Regulation B, a set of rules that lenders must adhere to when obtaining and processing credit information, “allows the use of advanced explainability techniques” associated with AI and ML-powered credit underwriting.

Agencies should also encourage government-sponsored enterprises, like Fannie Mae and Freddie Mac , and the Federal Housing Finance Agency to study and approve alternatives to traditional credit scores, Zest AI asserted.