LN: Shao et al. (2025) — Future of Work with AI Agents: Auditing Automation and Augmentation Potential

Bibliographic Reference

Citation: Shao, Y., Wang, J., Lam, J., et al. (2025). Future of work with AI agents: Auditing automation and augmentation potential across the U.S. workforce. arXiv:2506.06576. https://arxiv.org/abs/2506.06576

Important Note

Overview

The bibliography entry lists “Sapkota, S., Hoda, R., & Hassan, A.” as authors and an inaccurate subtitle (“Auditing Automation and Readiness”). The verified first author is Yijia Shao (Stanford). The correct title includes “across the U.S. Workforce.”


Pass 1 — Bird’s Eye View (5 Cs)

CAssessment
CategoryEmpirical study + measurement
ContextLarge-scale workforce automation potential study using “WORKBank” task database
CorrectnessSystematic methodology across occupations. Well-documented.
Contributions(1) WORKBank: database of O*NET tasks mapped to AI automation potential; (2) Project manager tasks ~40% automatable with current AI; (3) Augmentation more likely than full automation for PM tasks
ClarityGood.

Relevance: ⭐⭐⭐

Provides empirical evidence for which PM tasks are automatable, supporting PUMA’s task selection (triage, estimation vs. full project management).


PUMA Connection

The ~40% automation estimate for PM tasks validates PUMA’s design choice: PUMA targets specific high-volume, low-complexity tasks (triage, estimation) not full PM automation. Reference for Section 1.1 and the ethics section (1.3).

MOCs