--- title: "Dull, Dirty, Dangerous? Working Conditions and the Limits of Occupational AI Exposure · AI Fault Line" url: https://aifaultline.com/research/p/title-ff7ea6303f0df3ad/ description: "AI Fault Line is a neutral, evidence-cited field guide to why experts looking at the same developments disagree about AI's effect on productivity, jobs, inequality, regulation and the pace of change." --- [Research](https://aifaultline.com/research/) # Dull, Dirty, Dangerous? Working Conditions and the Limits of Occupational AI Exposure [Siyi Yu](https://aifaultline.com/authors/siyi-yu/) · [arXiv](https://aifaultline.com/sources/arxiv/) · 4 Nov 2026 [Read original source](https://arxiv.org/abs/2611.06888v1) Open access Where this sits in the debate - [Will AI destroy more jobs than it creates?](https://aifaultline.com/debate/jobs/) Direction of the finding as read from the abstract by an AI model ([AI-written](https://aifaultline.com/about/)). ## Abstract Automation is often justified as relief from dull, dirty, and dangerous work. I examine how these conditions relate to occupational AI exposure using 21 O*NET items across 754 US occupations. Higher exposure is associated with lower condition scores under a large language model (LLM) measure and AI Occupational Exposure (AIOE), with equal occupation weights and employment weights. Adjustment for physical demands, required education, training and experience, and wages reduces the composite association. The adjusted confidence intervals include zero for LLM exposure and remain below zero for AIOE. Separating repetition from restricted discretion also changes what the combined Dull score conveys. These comparisons show how item groupings and occupational characteristics shape the interpretation of exposure estimates. Occupational profiles offer a starting point for selecting contrasting research settings. Establishing desired automation or realized relief requires evidence about workers' preferences and specific interventions. ## About this publication **Preprint.** Posted on a preprint server or repository (arXiv, SSRN, Zenodo and others) before or without peer review. Anyone can post there. [How sources are ranked](https://aifaultline.com/about/#source-ranking). Published 2026 (recently) Citations Not recorded for this source Citation counts come from OpenAlex; none are recorded for this source — absence of a count is not absence of citations. [Chart: published 2026] ## Related papers On the same debate question, key papers first; related by topic, not by method or conclusion. Browse all: [Jobs](https://aifaultline.com/research/?question=jobs). - [Thousands of AI Authors on the Future of AI](https://aifaultline.com/research/p/doi-10-1613-jair-1-19087/) Katja Grace et al. · Journal of Artificial Intelligence Research · 2025 - [How Might Fiscal Policy Respond to the Rise of Artificial Intelligence?](https://aifaultline.com/research/p/doi-10-3386-w35437/) Karen Dynan et al. · NBER · 2026 - [Forecasting the Economic Effects of AI](https://aifaultline.com/research/p/title-1291a6248b6de2d6/) Ezra Karger et al. · NBER · 2026 - [The future of employment: How susceptible are jobs to computerisation?](https://aifaultline.com/research/p/doi-10-1016-j-techfore-2016-08-019/) Carl Benedikt Frey and Michael A. Osborne · Technological Forecasting and Social Change · 2016 Overview assembled automatically from collected metadata. [How this site is built](https://aifaultline.com/about/).