Artificial Intelligence and the Transformation of Work : A Multidimensional Framework for Assessing Technological Exposure, Adoption, and Employment Impacts

Zen Revista · Zenodo (CERN European Organization for Nuclear Research) · 15 Oct 2026

Where this sits in the debate

Will AI destroy more jobs than it creates? · its finding: it depends (why studies differ)

Direction of the finding as read from the abstract by an AI model (AI-written).

Key takeaway AI-written

A thesis that separates four constructs often conflated in jobs-at-risk estimates: digital AI exposure, physical automation, actual adoption, and net employment impact after displacement and new-task creation. It formalises these as sector-level indices and applies the framework to assess AI's employment consequences across sectors.

The paper in three steps AI-written

Written by an AI agent from the paper's abstract, which is the primary source and is quoted below.

The question

How can the employment effects of AI and automation be assessed without collapsing them into one jobs-at-risk number?

The approach

A thesis that defines four separate measures: digital AI exposure, physical automation potential, actual adoption, and net employment impact after displacement and new tasks. It applies them to nine sectors using about 280 published sources.

The findings

The abstract states the framework is formalized with explicit index definitions and uncertainty treatment and applied across nine sectors from IT to retail; it argues that single-number claims conflate different technologies, tasks and adoption conditions.

How to read it: review. A synthesis of existing studies — breadth over novelty; quality depends on what it includes and excludes.

Limitations and counterpoints AI-written

The abstract does not report the sector results, and the framework draws on secondary sources rather than new data.

Abstract

Artificial intelligence (AI) and automation are transforming the productive core of the global economy, yet public debate remains dominated by single-number claims about the proportion of jobs “at risk”—claims that collapse heterogeneous technologies, tasks, and adoption conditions into a single misleading index. This thesis develops, verifies, and applies a multidimensional framework for assessing the employment consequences of AI and automation across economic sectors. The framework distinguishes four constructs that previous research has often conflated: (D1) digital AI exposure, defined as the proportion of a sector’s task content that current AI systems can technically support or perform; (D2) physical automation, defined as the automatable share of embodied, manual tasks; (D3) adoption, defined as the actual, observed deployment of AI and automation technologies within a sector; and (D4) net employment impact, defined as the estimated medium-term effect on total employment after accounting for displacement, reinstatement, and the creation of new tasks. The thesis proceeds in four stages. First, it reconstructs the technical foundations of machine learning, generative models, and robotics with sufficient rigor for economic analysis, while tracing the economic theory of technology-induced labor reallocation from Ricardo’s machinery question to the contemporary task-based literature. Second, it formalizes the four-dimensional framework through explicit index definitions, aggregation functions, and uncertainty treatment, deriving each dimension from documented methodologies, including task-exposure assessments, adoption surveys, and causal labor-market studies. Third, it applies the framework to nine sectors: information technology, finance and insurance, healthcare, education, manufacturing, construction, transportation and logistics, hospitality and food services, and retail and e-commerce. This analysis synthesizes approximately 280 verified sources, including

About this publication

Preprint. A preprint — posted publicly before or without journal peer review; findings may still be revised.

Published: 2026 (recently)
Citations: 0 (OpenAlex)