Federal Reserve Governor Christopher Waller stated on Tuesday that artificial intelligence can enhance the efficiency of the central bank's internal operations but must be accompanied by stringent safety measures. He made these remarks at a European Central Bank research conference in Frankfurt.

Robust Safeguards to Ensure Data Security Waller noted that the Federal Reserve has established "strict guardrails" when using AI to protect sensitive policy information. "In research, you have to put guardrails on the use of AI. We probably wouldn't let you run AI on a laptop on an airplane that has sensitive policy information on it," he said. He added in a semi-joking tone, "They don't even let me use AI to clean up my emails."

Waller revealed that the Fed accesses various AI models through Amazon Web Services but, for security reasons, the versions used may not be the most cutting-edge technology. "We probably don't have the latest, most cutting-edge models because we need to do security checks and put in place safeguards," he said.

Practical Applications of AI Despite these limitations, Waller described AI as an "amazing tool" and listed specific use cases within the Fed, including summarizing meeting minutes. He pointed out that processing meeting records through AI can "very quickly and efficiently produce an executive summary of the minutes."

Waller emphasized that AI is positioned to assist humans, not replace them. He compared the current development of AI to the introduction of ATMs, which did not eliminate bank teller jobs but redirected human effort toward "higher-value activities." "I don't think you're going to take humans out of the picture and have AI do everything and we all go work at McDonald's," he said.

Ongoing Monitoring of External Application Risks Beyond internal use, the Federal Reserve is also closely monitoring how businesses apply AI technology in areas such as credit approval and compliance, as well as its potential impact on the labor market and monetary policy. Waller has previously stated that the Fed's application of AI is "not optional" and that AI literacy has been incorporated into employee performance goals to ensure the technology becomes a "lasting part of the Fed's operations."