CNB seminar "Generative AI as Routine-Biased Technical Change? Evidence from a Field Experiment in Central Banking"
Prague, 9 April 2025
Aleš Maršal (National Bank of Slovakia)
Aleš Maršal is a Senior Research Fellow at King Abdullah Petroleum Studies and Research Center in Saudi Arabia. He is at a long term sabbatical at National Bank of Slovakia. He is active in Czech academia and part of the SYRI project at CERGE-EI. Previously, he was teaching at MENDEL University in Brno (finance) and University of Economics in Prague (PhD level macroeconomics).
Generative AI as Routine-Biased Technical Change? Evidence from a Field Experiment in Central Banking (abstract)
Theories of routine-biased technical change posit that technological advancements complement workers in non-routine tasks and substitute for workers in routine tasks. We examine whether generative AI exhibits characteristics of routine-biased technical change through a field experiment at the National Bank of Slovakia. We randomly assign generative AI access to central bank employees completing workplace tasks that mirror the theoretical task-based framework. Our results indicate that generative AI access leads to large improvements in both quality and efficiency for the vast majority of participants. In line with theories of routine-biased technical change, we find that the benefits of generative AI are substantially larger in non-routine versus routine tasks. We also find some support for generative AI as cognitive-biased technical change, though smaller in magnitude than our tests of routine-biased technical change. While workers in routine intensive jobs see larger individual performance gains, GAI is less effective for the routine task content of their work, revealing a potential mismatch between worker-level and task-level impacts. Additionally, we find important differences in which workers benefit from generative AI. Low productivity workers benefit most in terms of quality while high-productivity workers benefit most in terms of efficiency. Our findings provide strong empirical evidence for generative AI as routine biased technical change, with important implications for how generative AI will impact workers and labor markets more broadly.