拥抱人工智能：意大利和美国家庭的比较

Embracing gen AI: a comparisonof Italian and US households by David Loschiavo, Olivier Armantier, Antonio DallaZuanna, Leonardo Gambacorta, Mirko Moscatelli andIlaria Supino January 2026 Keywords: generative artificial intelligence, technologyadoption, cross-country comparison, sociodemographic BISWorking Papers are written by members of the Monetary and EconomicDepartment of the Bank for International Settlements, and from time to time by othereconomists, and are published by the Bank. The papers are on subjects of topical This publication is available on the BIS website (www.bis.org). Embracing GenAI: A comparison of Italian and US households David Loschiavo1, Olivier Armantier2, Antonio Dalla Zuanna1, LeonardoGambacorta3, Mirko Moscatelli1, and Ilaria Supino1Bank of Italy2Economic Science Institute - Chapman University3Bank for International Settlements and CEPR Abstract This paper explores the household adoption of Generative Artificial Intelligence (GenAI) inthe United States and Italy, leveraging survey data to compare usage patterns, demographicinfluences, and employment sectoral composition effects.Our findings reveal higher adoption 1Introduction The development of Generative Artificial Intelligence (GenAI), particularly since the introductionof ChatGPT in late 2022, has marked a turning point in technology. GenAI can perform complextasks, from generating human-like text to assisting with creative, educational, and professional Several studies have documented increases in productivity associated with the adoption ofGenAI (Brynjolfsson et al., 2023, Noy and Zhang, 2023, Peng et al., 2023, Cui et al., 2024, Dell’Acquaet al., 2025).Despite its transformative potential, however, the adoption of GenAI varies signifi-cantly across countries, reflecting differences in digital infrastructures, income levels, human capital,cultural attitudes, and technological readiness (Liu et al., 2025, Calvino and Fontanelli, 2023). Un- This paper is the first to compare GenAI adoption among households in two advancedeconomies, drawing on comparable survey data from the United States and Italy.The two coun- We find that GenAI usage, whether one-off or more regular over the past year, is higherin the U.S. than in Italy. The difference is entirely driven by the heterogeneous socio-demographiccomposition. Looking into factors potentially explaining current and future use of GenAI tools, we ology used in the analysis. Section 3 examines adoption rates and their demographic and sectoraldeterminants.Section 4 explores differences in perceptions and trust in GenAI between the two 2Data We use data from two surveys, the Survey of Consumer Expectations (SCE) carried out by theFederal Reserve Bank of New York in February 2024 and the Household Outlook Survey (HOS)conducted by the Bank of Italy in August and September 2024.The two surveys share a set ofsocio-demographic variables (age group: under 40 years, between 40 and 60 years, over 60 years; employment status:employed, retired, other non-employed; gender; education:holding a collegedegree) and a special GenAI module. The module, addressed only to respondents and not to entire To conduct our analysis, we merge the two surveys into a single dataset using the populationweights provided by each2. We further calibrate the weights so that the proportions of respondents byage group, gender, education and employment status match the official data on the adult populationprovided by the United States Census Bureau and the Italian National Institute of Statistics3. 2.1The Survey of Consumer Expectations (SCE) The SCE is a monthly, internet-based survey produced by the Federal Reserve Bank of New York.Launched in 2013, it has become a valuable resource for researchers and policymakers to understand dents are drawn each month to match demographic targets from the American Community Survey,and they remain in the panel for up to 12 months before being replaced. 2.2The Household Outlook Survey (HOS) The Household Outlook Survey4 Italy since 20245.To meet Bank of Italy’s informational needs, the HOS is designed to track the evolution of households’ economic conditions in the years when the main Bank of Italy’s Survey onHousehold Income and Wealth (SHIW) is not conducted. Although the HOS is carried out online, the target sample is selected from the respondents to the previous SHIW wave which is insteadperformed through in-person interviews6.This sampling approach offers two key advantages:i) 3The adoption of GenAI in the United States and in Italy First, we investigate differences in the proportion of people that have used GenAI at least once in theprevious year (generic use); then, we focus on the differences in the share of people that have usedGenAI at least once a week over the past 12 months (regular use). The distinction between regular Based on the results of the SCE conducted in February 2024, the share of U.S. peoplereporting generic GenAI use was 36.4 p