AI agents that move idle cash into better-paying accounts could return billions to savers, but leave banks facing higher interest costs. A scenario reported by Korea JoongAng Daily puts a striking figure on the trade-off: shifting US$3.8 trillion in US savings deposits to accounts paying 4% could mean US$79 billion more in annual interest costs for banks.
Watch Desk analysis
What happened
Korea JoongAng Daily explores what may change as AI moves from answering questions to acting for people. An agent with permission to compare financial products and move money could seek higher returns on a customer’s behalf, rather than relying on them to notice a better rate and switch.
The article cites a scenario in which moving US$3.8 trillion in US savings deposits to accounts paying 4% would add US$79 billion in annual interest costs for banks. It says that amount could wipe out the net income of 32 banks; a more aggressive scenario, keeping only the minimum in checking accounts, could put as many as 78 banks in the red. These are estimates, not a forecast of what AI agents will actually move.
The piece also points to South Korea’s MyData service, which allows designated providers, with customer consent, to use financial information to request lower loan rates. The article says 162,000 people had benefited by July, saving a combined 100.3 billion won. It says the service is expected to expand to other financial tasks.
Why it matters
Better rates could mean real money back in savers’ pockets. But moving deposits at scale could also raise banks’ funding costs, and that may affect how they compete for customers. The same automation that helps a person escape a low-interest account could make consumer inertia a less reliable business strategy.
There is a difference between an agent suggesting a better deal and one authorised to transfer money or sign a customer up. The article cites a survey in which only 7% of respondents said they would allow AI to move money between bank accounts. It also notes that identity checks, payment rules and financial-consumer protections may limit what agents can do.
Our read
The useful question is not just whether an AI can find a better rate. It is whether people can set clear limits, understand the deal being recommended and stay in control when an agent can act. The reported savings through MyData show that financial automation can have a practical payoff; the larger US figures are a scenario, not evidence that a bank run by bots is around the corner.
What to watch
- Whether financial services let AI agents compare products and move money, and on what terms.
- What permissions, identity checks and consumer protections apply before an agent can act.
- Whether better automated rate-shopping changes banks’ offers or deposit costs.
- Whether MyData’s planned expansion produces further reported savings for customers.
Discussion spark: If an AI agent could move your savings to a better rate, what should it be allowed to do without asking you again?
Sources and evidence
- Source update (12 October 2026, 22:01 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.