Could AI agents break fractional-reserve banking?

Banks rely on a delicate balance of customer deposits for short-term withdrawals and long-term loans. This system works because most balances remain stable as customers rarely withdraw all their funds at once. However, AI agents could disrupt this equilibrium by constantly seeking better returns, potentially leading to faster bank runs. An AI agent could quickly transfer money to higher-yielding accounts, causing instability.

The real threat may not be bank runs but rather banks having to compete harder for funding. Customers value more than just interest rates, such as payment services, cash access, and customer service. AI agents could help customers better understand their needs and direct funds accordingly, moving surplus money to accounts with better yields.

This could put pressure on banks to pay for stability, either through higher savings rates or by offering fixed-term deposits with guaranteed returns. AI agents could help banks identify which customers can afford to keep funds for longer periods, creating a new revenue stream.

However, there is a risk of cascading withdrawals if millions of AI agents use similar decision-making rules. This could turn a minor pricing issue into a liquidity crisis, as banks would have to replace funding at significantly higher costs. Banks may choose to pass the increased costs to borrowers through higher interest rates on loans, but this is not guaranteed.

In the long run, AI agents could make it easier for customers to compare loan offers, apply for loans, and identify optimal refinancing opportunities. This could put pressure on banks to remain competitive in loan pricing, as borrowers may become more price-sensitive. The impact of AI agents on the banking industry remains uncertain, but their potential to change consumer behavior and reshape funding dynamics cannot be ignored.