Kim et al., "The Potential Impact of Disruptive AI Innovations on U.S. Occupations" (2025)

Munjung Kim, Marios Constantinides, Sanja Šćepanović, Yong-Yeol Ahn, and Daniele Quercia, Submitted (2025)
arXiv

@article{kim2025potential,
    author = {Munjung Kim and Marios Constantinides and Sanja Šćepanović and Yong-Yeol Ahn and Daniele Quercia},
    title = {The Potential Impact of Disruptive AI Innovations on U.S. Occupations},
    year = {2025},
    eprint = {2507.11403},
    archivePrefix = {arXiv},
    primaryClass = {cs.CY},
}

We analyze 3,237 AI patents (2015–2022) and distinguish “consolidating” AI that reinforces existing structures from “disruptive” AI that transforms them. Consolidating AI targets physical, routine, and solo tasks common in manufacturing and construction, while disruptive AI affects unpredictable and mental tasks in coastal science and technology sectors. Disruptive AI disproportionately impacts regions already experiencing skilled labor shortages, suggesting these technologies accelerate change where workers are scarce rather than simply replacing surplus labor.