Dull, Dirty, Dangerous? Working Conditions and the Limits of Occupational AI Exposure

A growing body of literature measures the extent to which occupations are exposed to AI, yet existing measures capture where AI could perform tasks rather than whether workers have actually adopted it. We introduce a distinct tier of exposure, delegated exposure, which records whether a worker has committed a task to AI by embedding it into a structured workflow. We operationalize this concept through the Agentic Adoption Index (AAI), measuring how closely an occupation's tasks align with the agentic routines that practitioners have built and shared. Using semantic embeddings of roughly 888,000 agent skill specifications from public GitHub repositories, we compute their similarity to nearly 18,000 O*NET task statements and aggregate these scores to the occupational level. We present three main findings. First, the occupations where task delegation concentrates differ sharply from those identified as most vulnerable by pre-AI automation frameworks. Second, the AAI aligns more closely with measures of technical capability than with measures of current conversational LLM use. Third, for occupations requiring a bachelor's degree or less, the AAI increases alongside average wage levels; however, this relationship reverses for occupations requiring a master's degree or higher, where adoption declines among higher earners. These patterns replicate on an independently collected corpus of agent skills from the Manus Skills Marketplace. This lower adoption among highly educated, high-earning workers may reflect tasks that inherently resist advance specification or professional discretion over the pacing of workflow codification. Distinguishing these mechanisms will require longitudinal measurement.

Exposure indices that map technical capabilities onto occupational abilities have become a standard tool for anticipating how AI may reshape work, but existing measures either cover software AI or proxy robotics through patents and installation data that reflect the industrial robots of past decades. We construct the first abilities-based exposure measures for robotics and physical AI, and the first to span occupations, industries, and geographies. Following the framework of Felten et al. (2018, 2021), we build a robotics capability taxonomy of 16 capabilities in eight domains from 78 academic, industry, and standards sources, use a large language model to score the relatedness of each capability to the 52 ONET work abilities, and aggregate to Robotics Occupational Exposure (ROE) for 774 occupations, Robotics Industry Exposure (RIE) for 250 NAICS industries, and county-level Robotics Geographic Exposure (RGE). We find that robotics exposure concentrates in physically intensive work such as construction, farming, transportation, and production, and in industries such as logging, mining, trucking, and warehousing. Unlike generative-AI exposure, ROE is higher in occupations with lower wages, lower education requirements, and less creative-ability content. Physical collaboration and assistance, terrain-based locomotion, grasping and handling, and task and functional reasoning move the exposure ranking most, yet no single capability carries the index. Because the measures capture technical relatedness rather than readiness, cost, or adoption, they should be read as an upper bound on where physical automation could affect work.

As Large Language Models reshape the global labor market, policymakers and workers need empirical data on which occupational skills may be most susceptible to automation. We present the Skill Automation Feasibility Index (SAFI), benchmarking four frontier LLMs -- LLaMA 3.3 70B, Mistral Large, Qwen 2.5 72B, and Gemini 2.5 Flash -- across 263 text-based tasks spanning all 35 skills in the U.S. Department of Labor's O*NET taxonomy (1,052 total model calls, 0% failure rate). Cross-referencing with real-world AI adoption data from the Anthropic Economic Index (756 occupations, 17,998 tasks), we propose an AI Impact Matrix -- an interpretive framework that positions skills along four quadrants: High Displacement Risk, Upskilling Required, AI-Augmented, and Lower Displacement Risk. Key findings: (1) Mathematics (SAFI: 73.2) and Programming (71.8) receive the highest automation feasibility scores; Active Listening (42.2) and Reading Comprehension (45.5) receive the lowest; (2) a "capability-demand inversion" where skills most demanded in AI-exposed jobs are those LLMs perform least well at in our benchmark; (3) 78.7% of observed AI interactions are augmentation, not automation; (4) all four models converge to similar skill profiles (3.6-point spread), suggesting that text-based automation feasibility may be more skill-dependent than model-dependent. SAFI measures LLM performance on text-based representations of skills, not full occupational execution. All data, code, and model responses are open-sourced.