The Speed of AI in Italy: Measuring the Pace of AI Penetration Across Occupations in Italy 2019-2025

Abstract

The recent advent of generative Artificial Intelligence and Large Language Models (LLMs) - technologies that enable the generation of multimodal content - is driving an increasingly broad penetration of technology into the realm of cognitive work. As a result, traditional labor market indicators are becoming less suitable for capturing the rapid evolution of the skills required to perform work activities. In this paper, we propose the application of the Task Exposure to AI (TEAI) index to the Italian labor market. TEAI is a measure designed to estimate the probability that occupational tasks are exposed to AI technologies. The index - constructed from the knowledge embedded in a selected set of LLMs - examines the evolution of task and occupation exposure to AI between 2019, the pre-LLM baseline year, and 2025, the post-adoption observation period. The analysis is conducted at the four-digit ISCO-08 occupational level (245 occupations surveyed), using Lightcast online job posting microdata for Italy as the source of labor demand information. This study makes three contributions: (i) it measures TEAI and its distribution across Italian occupations that, as observed in online job advertisements, show significant demand during the reference years; (ii) it introduces and applies the concept of Speed of AI, defined as the percentage change in TEAI between 2019 and 2025, in order to identify occupations with higher or lower rates of exposure growth within the period considered; and (iii) it provides a multidimensional map of the Italian labor market in which the initial level of exposure and the speed of change are jointly analyzed, offering policymakers and practitioners a monitoring tool with a strong policy orientation.