In a recent commentary, Torsten Slok, the chief investment strategist at Apollo Global Management, sounded an alarm about the potential for artificial‑intelligence‑powered agents to disrupt the traditional banking system. While AI has been hailed for its ability to streamline financial decision‑making and improve efficiency, Slok points out a darker side: the capacity of autonomous algorithms to rapidly move large sums of money from low‑interest, readily accessible checking accounts into more lucrative, higher‑yield options such as money‑market funds, short‑term bonds, or even emerging digital‑asset platforms. At the heart of the concern is the concept of a "bank run"—a scenario in which a sudden, mass withdrawal of deposits forces a bank to liquidate assets at fire‑sale prices, potentially jeopardizing its solvency. Historically, bank runs have been triggered by panic, rumors, or a loss of confidence among depositors.
Slok suggests that AI agents could create a new, technologically driven variant of this phenomenon. Because these agents operate on pre‑programmed criteria—such as maximizing net interest margin or minimizing exposure to low‑yield holdings—they can execute trades in fractions of a second, far faster than any human could. If a sizable portion of the population employs similar AI tools, the cumulative effect could be a swift, coordinated outflow of cash from traditional checking accounts. To understand why this matters, it helps to recall how banks traditionally fund themselves.
The majority of a bank’s capital comes from deposits, which are typically low‑cost liabilities. Customers keep money in checking accounts because of the convenience and the perception of safety, even though these accounts often earn little to no interest.
Banks then lend out these deposits at higher rates, earning a spread that underpins profitability. When AI agents systematically identify and redirect these low‑yield deposits toward higher‑return alternatives, the banks lose a crucial source of cheap funding. The resulting funding squeeze can force banks to raise deposit rates, tighten lending standards, or seek more expensive wholesale funding—all of which can compress margins and increase risk.
Slok emphasizes that the risk is not merely theoretical. The financial industry is already witnessing a surge in fintech applications that incorporate AI to optimize personal finance.
Robo‑advisors, for example, automatically rebalance portfolios, while AI‑driven budgeting apps suggest the best places to park idle cash. As these tools become more sophisticated, they begin to incorporate real‑time market data, predictive analytics, and even sentiment analysis to anticipate where the next yield advantage lies. When a large user base adopts such technology, the collective action resembles a coordinated strategy, albeit one executed by countless independent algorithms rather than a single orchestrated entity. There are also regulatory implications.
Traditional banking regulations—such as reserve requirements and deposit insurance—are predicated on the assumption that deposit withdrawals occur at a relatively predictable pace. AI‑induced rapid withdrawals could outpace the mechanisms designed to protect liquidity, potentially prompting regulators to intervene with emergency measures. Moreover, the opacity of algorithmic decision‑making makes it difficult for supervisors to anticipate or monitor systemic risk buildup. Slok calls for greater transparency in how AI agents are programmed to manage cash, as well as for stress‑testing frameworks that incorporate scenarios of mass algorithmic outflows.
From a macroeconomic perspective, the shift of funds away from checking accounts could also affect monetary policy transmission. Central banks influence the economy largely through the interest rates they set on short‑term government securities, which in turn affect the rates banks offer on deposits and loans. If a substantial share of deposits migrates to non‑bank instruments that are less directly tied to central‑bank policy rates, the effectiveness of monetary policy could be diluted. This decoupling might lead to unintended consequences, such as reduced credit availability for households and businesses that rely on traditional bank financing.
In response to these emerging challenges, Slok recommends several proactive steps. First, banks should invest in their own AI capabilities to retain customers by offering competitive, dynamic deposit products that adapt to market conditions in real time. Second, financial institutions need to enhance communication with clients, educating them about the trade‑offs between yield and safety, especially in an environment where higher returns often come with greater risk.
Third, regulators should consider updating liquidity coverage ratios and stress‑testing protocols to reflect the speed and scale at which AI agents can move capital. Finally, industry collaboration is essential: banks, fintech firms, and AI developers must work together to establish standards that balance innovation with systemic stability. In summary, while AI agents hold great promise for personal finance optimization, they also pose a novel systemic risk to the banking sector.
By automatically reallocating low‑interest deposits to higher‑yield alternatives, these algorithms could trigger a modern, technology‑driven bank run, undermining the cheap funding that underpins bank profitability and potentially weakening the transmission of monetary policy. Slok’s warning serves as a call to action for banks, regulators, and technology providers alike to anticipate these dynamics, adapt their strategies, and safeguard the stability of the financial system in an increasingly algorithmic world.