In a recent commentary, Torsten Slok, chief economist at Apollo Global Management, sounded an alarm about the emerging risk that artificial‑intelligence‑powered agents could inadvertently trigger a rapid outflow of funds from traditional bank deposits. While AI technologies promise to streamline personal finance management and improve the efficiency of savings decisions, Slok points out that these same tools could also accelerate the migration of cash from low‑interest checking accounts to more lucrative investment vehicles, potentially creating a self‑reinforcing cycle that resembles a modern‑day bank run.
At the heart of the concern is the way AI agents are being integrated into everyday financial workflows. Many fintech platforms now offer automated budgeting assistants, robo‑advisors, and cash‑management bots that continuously monitor a user’s account balances, spending patterns, and market conditions.
When these agents detect that a checking account is earning a rate that lags behind the returns offered by money‑market funds, short‑term Treasury bills, or high‑yield savings products, they are programmed to recommend, and in some cases automatically execute, the transfer of excess cash to those higher‑yield options. The logic is straightforward: maximize the return on idle cash.
However, Slok warns that when millions of households adopt such agents, the aggregate effect could be a sudden and massive shift of deposits away from banks that rely on low‑cost checking balances to fund loans and other credit activities. Traditional banks have long depended on a stable base of retail deposits to provide cheap funding for mortgages, consumer credit, and small‑business loans. Even though the interest rates on checking accounts are modest, the sheer volume of these deposits creates a reliable source of liquidity. If AI agents begin to reallocate even a modest portion of that liquidity—say, 5 to 10 percent—on a near‑real‑time basis, the cumulative impact could be substantial.
Slok illustrates this with a hypothetical scenario: if 100 million households each move $1,000 from a checking account to a money‑market fund within a short window, banks would see a $100 billion reduction in deposits almost overnight. Such a rapid contraction could force banks to sell assets at unfavorable prices, tighten credit, or raise funding costs, thereby amplifying financial stress across the system. The speed and automation inherent in AI agents differentiate this risk from traditional bank runs, which historically relied on human panic and the physical act of withdrawing cash. In the digital age, a “run” can be executed with a few clicks or even a single command to an algorithm.
The agents can monitor interest‑rate differentials across dozens of platforms, execute transfers instantly, and do so without the emotional hesitation that might slow down a human depositor. This creates a feedback loop: as more cash leaves the banking system, the remaining deposits become a smaller, less stable base, prompting banks to raise rates on checking accounts to retain customers. Higher rates, however, erode the banks’ net interest margins, potentially leading to tighter lending standards or higher loan rates, which in turn could dampen economic activity. Slok also emphasizes that the risk is not limited to the United States.
In many emerging markets, where banking infrastructure is less mature and interest‑rate spreads are wider, AI‑driven cash‑management tools could have an even more pronounced effect. In those economies, a sudden outflow of deposits could destabilize banks that already operate with thinner capital buffers, increasing the likelihood of sovereign or systemic crises. Regulators are beginning to take note.
The Federal Reserve and other supervisory bodies have started to examine the systemic implications of fintech and AI integration. Slok suggests that policymakers should consider a framework that encourages transparency in how AI agents make allocation decisions, as well as safeguards that limit the speed or volume of automated transfers during periods of market stress. Potential measures could include caps on daily transfer amounts, mandatory cooling‑off periods before large moves can be executed, or the requirement for agents to disclose the underlying assumptions that drive their recommendations.
From a consumer perspective, the benefits of AI‑assisted financial management are undeniable. Households can achieve higher returns on idle cash, reduce the friction of manually moving money, and benefit from data‑driven insights that help them meet savings goals.
However, Slok warns that the collective behavior of millions of such agents could unintentionally create a macro‑level vulnerability. He calls for a balanced approach that preserves the efficiency gains of AI while instituting prudent oversight to prevent a cascade of deposit withdrawals that could strain the banking sector. In summary, Torsten Slok’s warning underscores a paradox of modern finance: the very technology designed to optimize personal wealth management could, if left unchecked, undermine the stability of the institutions that underpin the broader economy. By recognizing the systemic implications of AI‑driven cash‑allocation tools and implementing thoughtful regulatory safeguards, policymakers can help ensure that the transition to a more automated financial ecosystem enhances, rather than threatens, overall financial resilience.