Future of Work in the Age of Automation, Augmentation, and Agentic AI (M-RCBG AWP 276)
B · Useful, limited A long Harvard Kennedy School senior-fellow working draft by practitioner Sol Rashidi that compiles others' labour-market evidence into new role frameworks and policy proposals; useful synthesis, not reviewed and with no new data.
Summary based on the full text·Original: about 3 hours to read
In short
A 97-page working draft (last updated 5 June 2026) from Sol Rashidi, Senior Fellow at Harvard's Mossavar-Rahmani Center (M-RCBG Associate Working Paper 276). Its thesis is that AI's main risk is a "talent formation fracture": automating entry-level work removes the path by which people become experts. It proposes an M-shaped talent model, an Orchestrator-Operator-Artisan (OOA) role typology, and a set of policy interventions.
Key numbers
- 6-13%employment decline for early-career workers in AI-exposed occupations, as cited from Stanford Digital Economy Lab
- 15-20%drop in graduate-level job postings, as cited from Handshake
- 14%decline in job-finding rates post-ChatGPT versus 2022, as cited from Massenkoff and McCrory
- 36O*NET tasks of a Search Marketing Strategist regrouped into four families in the worked example
As reported by the source.
English·PDF·opens hks.harvard.edu
What it reports
- Cited evidence on early careers: 6-13% employment decline for workers aged 22-25 in exposed jobs, 15-20% fewer graduate postings, 14% lower job-finding rates, recent-graduate unemployment at 5.6%.
- Argues the T-shaped talent model fits stable environments only and proposes an M-shaped model with several depths plus AI judgment.
- OOA typology: Orchestrators design and govern workflows, Operators supervise AI-mediated execution, Artisans supply deep expertise and tacit judgment. Archetypes describe modes of work, not job titles.
- Worked example: Search Marketing Strategist (SOC 13-1161 category, about 941,700 US workers, median wage $76,950), with its 36 O*NET tasks scored against the Eloundou exposure rubric.
- Policy proposals include an "AI Nutrition Label", an AI agent payroll tax above a workforce-impact threshold, early-career hiring tax credits, AI deployment registries, AI-ISO standards, extended antitrust, displacement notification and a capability "kill switch".
- Introduces a trademarked term, "Intellectual Atrophy", for cognitive offloading risk.
How they got it
Literature review across labour economics, cognitive science and management research, plus framework design. Appendices give a pay-structure method and task-scoring sensitivity analysis for the worked example.
Limits
- No original dataset; headline numbers are secondary and some mix sources and periods.
- Working draft, explicitly not formally reviewed.
- Policy section counts seven interventions in the abstract and eight in the contents.
- Trademarked concepts and framework branding point to a consulting audience.
Why it matters for companies
The question it raises is practical for workforce planning: if AI does junior work, where do future seniors learn judgment? The OOA typology gives HR and line leaders a vocabulary for redesigning roles by task rather than by title. Use the cited numbers through their original studies.
Summary based on the full text. Read the original for the full argument and data.Open PDF on hks.harvard.edu
- Tags
- Future of work, AI agents