The Bank of England (BoE) is actively experimenting with artificial intelligence (AI) to enhance its capabilities in forecasting inflation and refining its communication strategies, according to rate-setter Megan Greene. Speaking at a conference in London on Thursday, Greene revealed that the central bank is exploring various applications for AI technology.

Greene detailed that the BoE is testing AI for several critical functions, including providing early warning signals for potential financial crises, analyzing the intricacies of the labour market, and generating useful short-term predictions regarding the direction of consumer prices. These experiments aim to leverage AI's analytical power for more informed economic assessments.

While Greene acknowledged "massive" opportunities in using AI to track the economy and saw clear use cases for the central bank's day-to-day operations, she also sounded a note of caution. She emphasized that the stakes are particularly high when central banks employ AI for crucial areas like monetary policy and financial stability efforts.

The BoE's adoption of AI coincides with significant internal developments, including an upgrade of its technology infrastructure and an overhaul of its forecasting and communications practices. This overhaul follows a review conducted by former Federal Reserve Chair Ben Bernanke. Furthermore, this exploration comes amidst recent turmoil at the Office for National Statistics (ONS), which undermined market confidence after the agency on Thursday said it overstated the pace of inflation in April due to erroneous information from the government's transportation department.

The central bank is currently undertaking a series of targeted AI experiments as it formulates a comprehensive strategy for integrating the technology. Illustrating the growing acceptance of AI, Governor Andrew Bailey suggested earlier this week that he had used AI to help condense his speeches.

Greene elaborated that BoE staff have employed "machine learning techniques" to assist in producing inflation forecasts. Officials found they are "more accurate at forecasting CPI at shorter time horizons so it can be useful there, particularly as cross checks to more traditional models," she added. Greene also said staff had used AI "to figure out how comms might be improved and picked up more in the media."

Despite the promising applications for operational tasks, Greene expressed skepticism about AI's potential to significantly boost the UK economy's anemic growth and productivity rates in the near term. "There is an argument that AI should boost productivity growth, and quite quickly. This could happen through income effects, so workers will earn more and will either work less and have more leisure time, or will work the same and just consume more," she said. "I have my doubts that we'll see real productivity gains over my forecast horizon, which again is two to three years."