Study: AI Automation Threatens Female-Dominated Clerical Jobs, Risks Deepening Gender Gaps

A new joint analysis from the Brookings Institution and the Centre for the Governance of AI, drawing on public labor records and proprietary Lightcast modeling, highlights concentrated automation risk in routine administrative and clerical roles. The study estimates over six million workers face particularly difficult transitions after displacement because of limited transferable skills, age profiles, and thin savings; women constitute the vast majority of that vulnerable cohort. Jobs such as secretaries, payroll clerks and municipal clerks are singled out because core tasks—scheduling, data entry, standardized document handling—are increasingly performable by AI-driven software. The authors locate this change in a multi-decade decline in demand for routine back-office labor, arguing AI accelerates rather than initiates that trend. The report also finds divergence within exposed occupations: roughly seven in ten workers in jobs flagged as AI-exposed can plausibly transition to roles preserving earnings, especially where analytical, managerial or creative skills are available. Independent macroeconomists caution that, to date, there is limited evidence of mass, instantaneous replacement because of adoption costs, regulatory uncertainty and operational complexity, even as firm-level layoffs and worker surveys show early signs of disruption and anxiety. Complementary analyses highlight a supply-side amplification: concentrated infrastructure spending—estimated at about $1.5 trillion in 2025 with a projected expansion near $500 billion the following year—funnels influence to a small set of providers and can lock institutions into dominant stacks. That concentration, and the relatively tiny funding for decentralized alternatives (around $436 million in 2024), raises systemic risks that policy must address alongside worker retraining. The Brookings/Centre study therefore pairs near-term labor-market prescriptions—publicly funded retraining, employer reskilling, safety-net reforms—with calls for broader industrial and competition policies to preserve contestability, portability and auditability of AI systems. For individual workers the practical advice remains the same: accumulate transferable skills, experiment with AI tools to augment productivity, and seek credentialed pathways into less automatable roles. Absent coordinated employer and public responses, the pattern of displacement risks concentrating harm on households that historically relied on clerical roles as stable on-ramps into the labor market.