The Bank for International Settlements published a working paper comparing generative artificial intelligence (GenAI) adoption by households in the United States and Italy using harmonised survey modules. It finds higher reported GenAI use in the US, but attributes the cross-country gap primarily to differences in socio-demographic composition rather than systematically higher within-group usage in the US. Based on the surveys, 36.4% of US respondents reported using GenAI at least once in the past year versus 31.0% in Italy, while regular use (at least weekly) was 13.7% versus 11.7%. A Blinder-Oaxaca decomposition suggests the 5.4 percentage point generic-use gap is more than fully explained by demographics (10.1 percentage points), with within-group differences offsetting it (-4.7 percentage points), and the 2.0 percentage point regular-use gap is entirely explained by demographics. Despite lower usage, Italian respondents report more optimistic expectations for GenAI’s effects on well-being and financial wealth and relatively higher trust in government and institutions to handle personal data, while respondents in both countries generally trust GenAI-based services less than human-operated alternatives, especially in banking and public policy.