Bank of America’s Brian Moynihan Is Making the Case for Slower AI

Artificial intelligence is beginning to show up in U.S. economic data, but not in the form of sweeping automation or sudden productivity shocks. In recent remarks, BofA CEO Brian Moynihan described AI as a growing contributor to economic output and corporate efficiency, one whose impact is emerging gradually through capital investment, internal workflow changes, and measured deployment.

Speaking on Bloomberg Television, Moynihan said Bank of America expects U.S. GDP to grow 2.4% in 2026, outpacing other developed economies. AI, he said, is part of that outlook, though not the dominant factor. “It’s not all attributable to AI,” Moynihan said. “But that’s having a marginal impact. It’s pretty strong.”

That view aligns with recent research from the Bank of America Institute, which has reported that technology- and AI-related capital expenditures, particularly in software and computing, have supported U.S. GDP growth over the past year. In an October report, the institute said investment in those categories helped underpin economic resilience, while early effects on employment remained limited.

At Bank of America itself, Moynihan’s comments point to an approach shaped control over speed.

“Augmented intelligence,” not automation

Moynihan frames the bank’s strategy around what he calls “augmented intelligence,” emphasizing systems designed to support human decision-making rather than replace it. “We’ll be applying more and more of automated intelligence, or augmented intelligence, as we call it, with a person using AI, using that to be more effective,” he said on Bloomberg Television. “That’ll affect all the businesses.”

The bank has chosen to deploy AI inside a regulated environment. Moynihan stressed that data quality and governance are prerequisites for scale. “You have to have your data perfect,” he said. “You have to have the controls around it so the customer’s getting a good answer.” He added that the bank cannot allow systems to operate without guardrails. “We just can’t let it run and have answers customers won’t have faith in.”

Those constraints help explain why Bank of America’s AI rollout has been incremental. “That’s why it’ll be a little slower build out, I think, than people see,” Moynihan said, “but a very relentless build out.”

The bank’s technology spending reflects that discipline. Bank of America spends roughly $13 billion annually on technology, with about $4 billion directed toward new initiatives and product development. Moynihan said each project is evaluated based on expected revenue impact, cost savings, or operational simplification, with returns assessed against the bank’s cost of capital.

Other large banks have described similar guardrails. JPMorgan Chase, for example, has said it is deploying generative AI tools across internal workflows while maintaining oversight and controls, focusing on productivity gains rather than consumer-facing experimentation.

The revenue impact customers don’t see

While consumer-facing AI tools often attract the most attention, Moynihan said the near-term payoff is occurring inside the bank. “In the very near term, it’s mostly about process engineering,” he said, describing how AI is embedded into specific workflows to reduce manual work.

Those internal tools are increasingly tied to revenue-generating activities. Bank of America uses AI to help relationship managers and advisors prepare for client meetings, synthesize account data, and turn around materials more quickly. Reuters has reported that these tools are allowing bankers to cover more clients by reducing preparation time and administrative overhead: an efficiency gain with clear revenue implications, even if the bank does not break out AI-driven revenue explicitly.

In wealth management, AI-generated market summaries combined with client portfolio data are being used to support more personalized advice. Executives have described these systems as augmenting advisors’ work rather than automating it, reinforcing the bank’s emphasis on assistive use cases.

Similar patterns are emerging elsewhere in the industry. JPMorgan told Reuters that AI tools helped advisors manage client relationships and sales activity during periods of market volatility, highlighting how revenue impact is often indirect and internal rather than visible to consumers.

As AI investment accelerates across the economy, concerns about overbuilding and misallocated capital have grown. Moynihan addressed those concerns by emphasizing the narrowness of current exposure. “We see it as being relatively limited because it’s a narrow group of companies, narrow group of spending, and the companies are spending it have a lot of money,” he said.

From a lending standpoint, he said Bank of America evaluates AI-related projects using traditional credit criteria, including leverage, contract duration, and the quality of counterparties committed to long-term usage. “We think about the tenant, for lack of a better term, the quality tenant, the length of that tenant,” Moynihan said.

That posture is consistent with broader industry assessments. S&P Global has said AI is likely to be an incremental force in banking rather than a destabilizing one, provided institutions manage costs and governance carefully.

For Moynihan, AI is a long-term infrastructure build. “We are after it every day,” he said. “But a lot of this is really fairly controlled right now to make sure that we can actually deliver what we say when we put it in.”