While there's been a lot of talk about AI regulation in Washington, the future of policy around the technology and, more specifically, its impact on banks, remains uncertain.
Banks can use AI in every line of business, including commercial, retail and trading. They're using AI throughout their infrastructures and back offices, too. So policy around the technology will impact governance, strategies and innovation capabilities. However, while policymakers like Senate Majority Leader Chuck Schumer, D-N.Y., Elizabeth Warren, D-Mass., and Sen. Josh Hawley, R-Mo., have offered broad comments and proposals to steer AI regulation, details have been light, and industry-specific guidance for
As they wait for new rules to take shape in laws and regulations, banks are developing risk controls in tandem with their AI strategies as the technology evolves, and the
"It's important to have an industry-focused approach because there are industries, like banking, that already have a history of complying with laws and regulations that are adjacent to AI," Miller said.
Last week, Schumer hosted major technology executives to discuss potential AI regulation in a closed-door meeting attended by Meta CEO Mark Zuckerberg; X, formerly known as Twitter CEO Elon Musk and OpenAI CEO Sam Altman, among other top names in tech.
Schumer, who has been an avid proponent of AI policy, proposed in June
Politicians' proposals for AI have ranged from amending Section 230 to strip immunity from AI companies in claims involving generative AI to creating a new regulatory body to monitor Big Tech companies like Amazon and Facebook.
Eyvonne Mallett, an attorney at Loeb & Loeb, said banks should not only consider future movement, but also how policies like anti-discrimination and data privacy rules can be applied to AI. Sector regulators like the
This week, CFPB director Rohit Chopra said in a release that creditors must provide specific reasons for why credit is denied, whether the decision was made with AI or not. The agency released a note cautioning lenders against relying on data taken from sources outside a consumer's credit application.
"The challenge right now is, there's a lot of talk about AI regulation, but there hasn't really been much traction or movement," Mallett said. "So sitting in the role of the bank, I would really be thinking about those policies and regulations that are already in place, and how AI impacts them."
Miller said policies that could enhance AI regulation for financial institutions include establishing a common lexicon of AI terms and functions, clarifying the documentation and controls to satisfy examiners and increasing transparency around
The ABA created a cross-functional AI working group that coordinates with banks of varying sizes to collect concerns and insights, and is in the early stages of working with policymakers on how AI rules could impact banks, Miller said.
"We've had a lot of interest and really positive engagement with policymakers," Miller said. "Part of that is helping them learn about what banks are doing, and how they've always done this in a way that's safe and sound."
Mallett added that many banks prefer to stay off the radar when collaborating with politicians and regulators, so they go through avenues like trade groups, unlike notoriously public Big Tech executives like Musk.
Kevin Allison, president and head of research at AI consulting firm Minerva Technology Policy Advisors, said in an email that banks haven't been very active in high-level AI regulatory conversations, but they should be. Allison said bank CEOs and boards should be more vocal about AI issues and applications to help educate policymakers on how companies are using the technology, and to reassure customers that they're implementing AI responsibly.
"AI regulation will affect banks in many ways," Allison said. "Banks are critical economic infrastructure, so they have a lot riding on governments striking the right balance on AI safety and national security. ... Banks also have an interest in governments putting in place smart policies to encourage responsible adoption of AI across the wider economy, because if AI-driven productivity gains materialize, that's good for the people and companies who bank with them."
A new model rates large banks on their efforts to develop and deploy artificial intelligence technology.
While some banks have been working with AI for more than a decade, others are new to the technology. JPMorgan Chase, which has used AI in various use cases for years, recently brought on hundreds of data scientists, ethicists, risk professionals and others
Wells Fargo is working to
Mallett said AI governance and deployment can be resource-intensive, especially for smaller banks that need to outsource more of the work. She added that the banks she's advising are using audits to monitor their AI strategies, including data management, explainability and human review processes. Miller said he thinks banks' risk management can be an example for other industries adopting AI.
"Banks have been leaders in the field of using artificial intelligence for quite some time now," Miller said. "They have a strong and mature risk management framework, they've deployed the technology in a variety of ways. That's something that they're used to doing, and the regulators are used to holding them accountable for that."
