WASHINGTON DC: A new academic study has revealed that political pressure can sharply influence deliberations within the U.S. Federal Reserve, even in a simulated setting.
Researchers at George Washington University conducted an experiment replicating a Federal Open Market Committee (FOMC) meeting using artificial intelligence agents designed to mirror real-life policy makers.
Each AI agent was modeled on members past policy decisions, biographies, and public speeches.
The simulation processed real-time economic and financial data before reaching rate-setting decisions.
The results, published on August 31 by academics Sophia Kazinnik and Tara Sinclair, showed that when political pressure was introduced, divisions deepened, and dissent became more frequent among the simulated board members.
This experiment demonstrates that the Federal Reserve is only partially insulated from political influence, the paper noted. External scrutiny has the power to shape internal decision-making, even in institutions bound by formal rules.
The model replicated the Fed’s July 2025 FOMC meeting.
While central banks are not yet ready to let AI determine monetary policy, many are integrating the technology into research and analysis.
-
The Federal Reserve has used generative AI to analyze FOMC minutes.
-
The European Central Bank employs machine learning to forecast eurozone inflation.
-
The Bank of Japan applies AI to assess price drivers, with recent research highlighting a shift from raw material costs to labor costs.
-
The Reserve Bank of Australia is testing an AI tool that generates summaries on policy-related questions.
Governor Michele Bullock of the RBA stressed on September 3: “We are not using AI to set policy. Instead, we see it as a tool to improve efficiency and amplify staff capacity in research and analysis.
According to the Bank for International Settlements (BIS), central banks see AI as strategically important for enhancing data analysis and decision support.
However, it warned in an April report that many are still in the early stages of adoption, facing challenges around governance, transparency, and ensuring access to high-quality data.
The study highlights both the promise and the risks of AI in financial governance. While it can deepen economic insight, the simulation shows that political pressure even in a controlled model can destabilize consensus within one of the world’s most influential financial institutions.
