Selected Work
This page is deliberately small. It shows a few pieces that best represent the current range of public work: diagnostic review, issue analysis, framework development, working paper, and longer argument.
You Agreed to the Rule. Did You Agree to the Experiment?
A working paper on ICML 2026 and the missing boundary between agreement to perform review duties under a rule and agreement to have those duties used as experimental conditions.
Best for: AI governance, research ethics, peer review, institutional design, and experimental-participation audiences.AI Deployment Risk Clarity Review
A public diagnostic sample for identifying verification, escalation, rollback, accountability, and user or frontline burden gaps before wider AI deployment.
Best for: AI governance, product risk, trust and safety, policy, customer-support automation.Fast Answers, Slow Responsibility
A CIEA-M1 sample testing whether AI customer support removes the real bottleneck or shifts it into risk classification, legal responsibility, human escalation, and frontline labour.
Best for: public-service design, AI deployment review, labour and accountability audiences.The Landlord at the End of the AI Boom
An essay on spatial gate rent and how AI-sector value becomes collectible through ownership of urban access.
Best for: AI political economy, housing, urban policy, labour, and editorial audiences.Capital Played Global. Labor Reads Global Too.
An analysis of how wage settlements, bonus formulas, and housing-credit benefits become compensation benchmarks before full institutions travel.
Best for: labour, compensation, industrial relations, and political-economy readers.Who Really Controls South Korea’s Public School Reform?
A public-contract reading of certification authority, legal jurisdiction, language priority, data, and what remains in public hands after public spending.
Best for: education policy, public contracting, institutional design, and journalism.Feasible-Set Governance
A framework page on how candidate universes are structured before bounded actors appear to make a choice.
Best for: governance research, platform analysis, selection systems, and concept-driven readers.