Bank of America and Citi Move to Restrict AI Agents as Banks Weigh "Smart Wealth" Risk

Bank of America and Citi are moving to restrict the use of AI agents, including in so-called "smart wealth" scenarios, according to a report carried by East Money. The step marks a tightening of banks' previously open stance toward automated agents.

Framed around why banks are starting to fear smart wealth management, the report points to a core question: when an AI agent can check accounts, carry out operations and even make wealth decisions on a user's behalf, financial institutions have to reassess where the risk and the responsibility sit.

The background is a period in which agents have moved deeper into personal finance, from bill management to investment assistance. Banks want the efficiency such tools promise, but they also face real constraints around compliance, fraud and liability.

The report's headline does not spell out the specific form the restrictions take, but "moving to restrict" is itself a signal: institutions are no longer treating direct agent access to accounts and funds as a default.

The change will travel upstream. AI wealth tools that depend on bank interfaces may need to redesign their permission models and user-authorization flows to keep operating in a more constrained environment.

For users, the near-term effect may be a narrower set of available "smart wealth" features, or more explicit authorization steps before an agent can act.

What to watch next is whether other banks follow with similar limits, and whether regulators set clear rules for the boundaries of AI agent permissions in financial services.