The growing adoption of artificial intelligence (AI) in the workplace has raised questions about how employer demand for different work activities is changing. We examine how the prevalence and content of work activities in online job postings change in association with activity-level measures of AI exposure, derived from Copilot and Claude usage. To enable systematic comparisons, we develop a pipeline that extracts the responsibilities from job advertisements and maps them onto O*NET's standardized taxonomy of Intermediate Work Activities (IWAs). We then apply this approach to 50,000 job advertisements from 33 firms between 2023 and 2025. From this analysis, we first find that explicit mentions of AI increased by 66%, although much of this growth occurs outside descriptions of the actual work activities associated with a job. Looking specifically at the advertised job responsibilities, activities with greater AI exposure tend to decline in advertised prevalence, but only mildly as a share of the total number of responsibilities. At the same time, more AI-exposed work activities increasingly are associated with explicitly AI-related responsibilities, suggesting that employers are looking to hire people who can perform these responsibilities with AI.
Jiale Liu Pennsylvania State University, Victor Bursztyn Adobe Systems, Lin Ai, Haoliang Wang Adobe Research, Sunav Choudhary Adobe Research, Saayan Mitra Adobe Research, Qingyun Wu
原文摘要
In open-ended domains, teams must reconcile diverse viewpoints to produce strong deliverables. Answer aggregation approaches commonly used in closed domains are ill-suited to this setting, as they tend to suppress minority perspectives rather than resolve underlying disagreements. We present TeamFusion, a multi-agent system designed to support teamwork in open-ended domains by: 1. Instantiating a proxy agent for each team member conditioned on their expressed preferences; 2. Conducting a structured discussion to elicit agreements and disagreements; and 3. Synthesizing more consensus-oriented deliverables that feed into new iterations of discussion and synthesis. We evaluate TeamFusion on two teamwork tasks where team members can judge how well their individual views are represented in team decisions and how consensually good the final deliverables are, finding that it outperforms direct aggregation baselines across metrics, tasks, and team configurations.
cs.CL · cs.MA
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表