In a recent commentary, Torsten Slok, the chief economist at Apollo Global Management, raised a flag about a new, technology‑driven risk to the stability of the banking system. He warned that sophisticated artificial‑intelligence agents—software programs that can autonomously manage personal finances—could inadvertently trigger a modern‑day bank run. The mechanism, he explained, is simple yet potentially powerful: these AI agents are programmed to optimise the return on a user’s cash holdings.
When they detect that a traditional checking account is offering a near‑zero interest rate, they will automatically move the money into alternative vehicles that promise higher yields, such as money‑market funds, short‑term bond ETFs, or even emerging digital‑asset platforms. While each individual transfer might be modest, the cumulative effect across millions of households could be dramatic, draining banks of the cheap, readily‑available deposits they rely on to fund loans and other core activities.
### Why AI Agents Matter The rise of AI‑powered personal finance tools has been swift. From robo‑advisors that rebalance investment portfolios to chat‑based assistants that suggest budgeting tweaks, consumers are increasingly handing over routine financial decisions to algorithms. These agents are designed to act in real time, constantly scanning market data, interest‑rate changes, and product offerings from a plethora of financial institutions. When a bank lowers its checking‑account rate—or fails to raise it in line with inflation—an AI agent can instantly flag the inefficiency and execute a transfer to a more lucrative alternative, often without the user needing to approve each move.
This level of automation, while convenient, creates a feedback loop: as more users’ cash exits low‑yield accounts, banks feel pressure to raise rates or risk losing a critical source of funding. ### The Traditional Role of Cheap Deposits Historically, banks have depended on low‑interest checking and savings accounts to provide a stable, low‑cost source of capital. These deposits are typically considered “sticky” because customers rarely move them unless faced with a compelling reason, such as a significant fee increase or a major service disruption.
The stability of these funds allows banks to lend to businesses and individuals at rates that exceed the cost of the deposits, generating profit and supporting economic activity. When the supply of cheap deposits shrinks, banks must either raise the rates they pay—compressing their net interest margin—or turn to more expensive wholesale funding sources, which can increase borrowing costs for consumers and slow credit growth.
### How an AI‑Induced Run Differs From Traditional Runs A classic bank run occurs when depositors, fearing insolvency, rush to withdraw cash en masse, often spurred by rumors or panic. The speed and coordination of modern AI agents could produce a similar outcome, but with a digital twist.
Rather than queuing at a teller, the agents execute electronic transfers in milliseconds. Moreover, the decision to move money is not based on fear of a bank’s collapse but on a rational, data‑driven assessment of return on cash. This rational‑driven exodus could be even more destabilising because it may happen before any actual liquidity strain appears, catching banks off‑guard and leaving little time for corrective action. ### Potential Systemic Implications If a significant portion of the banking sector’s low‑cost deposits were to migrate to alternative platforms, several systemic risks could emerge: 1.
**Liquidity Pressures**: Banks could experience sudden shortfalls in cash, forcing them to sell assets at unfavorable prices or rely on central‑bank facilities, which may be costly or limited. 2.
**Interest‑Rate Competition**: To retain deposits, banks might feel compelled to raise rates on checking accounts, eroding their profit margins and potentially leading to higher loan rates for borrowers. 3. **Regulatory Challenges**: Regulators would need to monitor not just the health of individual institutions but also the aggregate flow of deposits driven by algorithmic decision‑making, a task that requires new data‑collection frameworks. 4.
**Market Concentration**: If alternative platforms—especially fintech firms or crypto‑based services—absorb large volumes of cash, the concentration of liquidity could shift away from the traditional banking system, altering the landscape of financial intermediation. ### Mitigation Strategies Slok suggested several avenues for policymakers and banks to mitigate this emerging risk: - **Transparency and Disclosure**: Banks could provide clearer, real‑time information about the rates they offer on deposit products, making it easier for AI agents to assess competitiveness without relying on third‑party data aggregators. - **Rate Floors or Incentive Programs**: Introducing minimum interest rates on checking accounts or offering tiered rewards for larger balances could help retain a core base of deposits.
- **Collaboration With Fintech**: Traditional banks might partner with fintech firms to embed AI‑friendly features within their own platforms, ensuring that any automated transfers stay within the bank’s ecosystem. - **Regulatory Oversight of Automated Transfers**: Regulators could consider guidelines that require AI‑driven financial tools to include safeguards—such as throttling the volume of transfers per day or requiring explicit user confirmation for large moves. ### The Bigger Picture The concern raised by Slok is emblematic of a broader shift in how technology is reshaping financial stability.
As algorithms become more capable of executing complex financial strategies at scale, the line between individual optimisation and systemic risk blurs. What begins as a benign effort to earn a few extra basis points on idle cash can, when multiplied across millions of accounts, generate macro‑level shocks. In conclusion, while AI agents promise greater efficiency and better returns for consumers, they also introduce a novel conduit through which capital can rapidly exit the traditional banking system. Stakeholders—including banks, fintech innovators, and regulators—must recognise this dual‑edge nature and work collaboratively to design safeguards that preserve the benefits of automation without compromising the stability of the financial system.
Torsten Slok’s warning serves as an early reminder that the integration of AI into everyday financial decisions is not just a convenience issue; it is a potential catalyst for systemic change that warrants careful attention and proactive policy design.