AI agents could trigger a bank run, says economist

AI agents could soon automate a portion of savings management, shifting deposits to the highest-yielding accounts. For Apollo, this shift could weaken the banking model by accelerating the flight of low-cost deposits. The article AI agents could trigger a bank run, says economist first appeared on Cryptoast.

The emergence of conversational agents equipped with autonomous decision-making capabilities could fundamentally reshape the dynamics of the traditional banking sector. According to a recent analysis by Apollo chief economist Torsten Slok, these intelligent assistants are now capable of proactively optimizing household financial management. By monitoring market opportunities in real-time, these automated tools could systematically shift client liquidity to accounts offering the highest yields, placing unprecedented pressure on legacy financial institutions.

The banking business model has historically relied on a behavioral flaw: saver inertia. Deposits in checking accounts, which offer very low interest, provide a cheap source of funding for banks. In the United States, this disparity is stark, with checking accounts offering yields near 0.1%, while secure alternatives provide returns between 3.3% and 5%. Should a critical mass of users delegate this management to algorithms, these tools would instantly eliminate such inertia, triggering a massive and automated outflow of deposits.

This phenomenon amounts to a new form of technological "bank run." Unlike classic liquidity crises sparked by irrational panic, this destabilization would be driven by the mathematical logic of profit maximization. Programmed to act simultaneously and identically the moment an opportunity arises, AI agents could drain commercial bank reserves in record time. Such a concentration of financial behavior threatens to squeeze net interest margins, forcing banks to drastically increase deposit rates to retain their customers.

The prospect of such financial uniformity was already anticipated by SEC Chair Gary Gensler as early as 2020. In an academic study, he highlighted that the widespread deployment of deep learning risked eliminating the diversity of market participant strategies, thereby threatening the overall stability of the system. Faced with this new landscape, regulators currently appear ill-equipped to oversee capital movements orchestrated by autonomous software. Banks must therefore rethink their deposit-gathering strategies to survive this technological acceleration of financial mobility.