AI gains raise or lower inflation depending on their timing
Bank of Italy research finds that AI adoption lowers inflation when productivity gains arrive quickly but raises it when gains emerge gradually. Its research newsletter also reports evidence on judicial productivity and financial vulnerabilities.
Productivity sets the inflation path
Lucia Esposito, Andrea Fabiani, Elia Moracci, Andrea Papetti and Massimiliano Pisani simulate AI adoption in a euro-area model.
Rapid productivity gains expand supply more than demand and lower inflation; gradual gains raise inflation.
Responding to inflation or the current natural rate of interest stabilizes prices across their scenarios.
A separate multi-country study by Anna Bartocci, Alessandro Cantelmo, Pietro Cova and Stefania Villa finds permanent GDP gains from AI adoption, with a small initial inflation increase as demand outpaces supply.
Euro-area effects are weaker with services-only adoption, smaller productivity gains than in the United States and China or greater precautionary saving.
Remote judging cuts court backlogs
Enrico Miglino studies 10 million Italian civil cases from 2016 to 2026.
Judicial productivity depends more on individual judges than on their courts, and higher output brings neither higher pay nor lower judgment quality.
Court transfers temporarily reduce productivity.
Over ten months, an incentive under Italy's National Recovery and Resilience Plan raised judges' productivity by one-sixth through remote work supporting overloaded courts.
Backlogs and disposition times in target courts both fell by 5 percent.
Scenarios are not forecasts
The AI studies make inflation outcomes conditional on how productivity gains develop.
Their results distinguish simulated mechanisms rather than establish a single inflation forecast.
The newsletter states that authors' views do not necessarily reflect the Bank of Italy's position.