In AI-simulated Fed meeting, political pressure polarises board

cross-posted from: https://sh.itjust.works/post/45841810

The findings showed when political pressure is applied, the AI agents became fragmented and dissent more common, according to the paper by Sophia Kazinnik and Tara Sinclair.

"This simulation shows that the Federal Reserve is only partially insulated from politics," and that "outside scrutiny can shape internal decision-making, even in an institution guided by formal rules," they said in the paper.

The simulation replicates the July 2025 FOMC meeting. While few central banks may be ready to have AI agents set monetary policy, a growing number of them are using the technology to streamline operations.

The Fed conducted research using generative AI models to analyse minutes of the FOMC meetings. The European Central Bank uses machine learning models to forecast euro-area inflation.