Get Full Government Meeting Transcripts, Videos, & Alerts Forever!
Get email alerts on the Artificial Intelligence topic
No spam. Unsubscribe anytime.
Panel: artificial intelligence and data reliability complicate Fed policymaking
Summary
Presenters warned that AI adoption, potential labor‑market disruption, and increasing unreliability of traditional and alternative data sources complicate monetary policy, urging data‑grounded approaches and caution when using machine‑learning-derived indicators.
Presenters at a Federal Reserve System briefing identified artificial intelligence and data‑quality challenges as central uncertainties for monetary policy and urged caution when incorporating alternative data and machine‑learning techniques.
A presenter (S1) summarized the problem: “The presence of artificial intelligence makes it important to have a data grounded approach, but increasing unreliability in data complicates this outlook.” They listed three key concerns: falling survey response rates (for example, the University of Michigan sentiment survey), budgetary constraints limiting BLS capacity and increasing reliance on imputed or predicted data, and noise in smaller‑sample alternative data (such as New York Fed surveys).
Presenters said AI poses multiple channels of risk: rapid, widespread adoption could produce faster job losses, wage polarization and sectoral price effects; slower, gradual adoption could be less disruptive. Panelists cited academic work that finds microeconomic productivity gains from AI may not translate into large aggregate labor‑market improvements, and noted a Fed study that found a positive correlation between a job’s AI exposure and unemployment.
On data and methods, presenters warned about the risk of algorithmic ‘hallucination’ when using machine learning on scraped or alternative sources and said ML tools must be carefully validated before being used for policy decisions. They pointed to retraining efforts noted in a New York Fed survey as an offsetting force that may limit widespread layoffs.
Implications for policy: Presenters said monetary policy can smooth short‑term transitions but is limited in addressing structural labor disruptions; fiscal and retraining policies will be important if AI causes persistent displacement. They recommended maintaining data‑dependence, broadening data sources cautiously and using ML techniques with rigorous validation.
What’s next: The panel urged monitoring both the pace of AI adoption and the reliability of data sources used for policymaking, including continued attention to survey response rates and budgetary constraints affecting official statistics.
