Where can AI be used? Insights from a deep ontology of work activities

arXiv. Evidence for decomposing work before judging where AI can be used.

Authors: Alice Cai, Iman YeckehZaare, Shuo Sun, Vasiliki Charisi, Xinru Wang, Aiman Imran, Robert Laubacher, Alok Prakash, Thomas W. Malone · Venue/status: arXiv · Year: 2026

This project builds a deep ontology of work activities and uses it to map where current AI software and robotic systems appear able to act. The paper argues that job-level categories are too coarse: we need activity-level structure before making claims about AI exposure, substitution, or complementarity. My role connects ontology construction, AI classification workflows, human aggregation, platform development, and the broader infrastructure question: how institutions can reason about future work at a useful level of detail.

The rendered paper page adds figures, authorship context, citation metadata, contribution notes, and related systems. Key links: home, systems, papers, manuscripts, Google Scholar, GitHub, LinkedIn, ORCID, MIT profile, and CV PDF.