Apollo chief economist Torsten Slok warns that AI agents designed to maximize returns could unintentionally help trigger bank runs. He says such agents may rapidly move household deposits from low-interest checking accounts into higher-yield alternatives when automated instructions are applied at scale.
The concern is that coordinated withdrawals could reduce a bank’s available liquidity, potentially accelerating customer fears. CoinDesk and Business Insider both report on Slok’s thesis that while consumers may use AI to optimize where their money earns interest, large, automated shifts could resemble the dynamics of a bank run. Neither report cites a specific incident or offers evidence that this is already happening; instead, they frame the discussion as a risk scenario tied to how AI-driven financial behavior could scale across many customers at once.