Human Agency
Judgment, problem formation, authorship, delegation, expertise formation, and the changing scale of individual agency.
Research
When does machine capability become human capability?
The Intellectual Agenda of AI Humanism studies how increasingly capable machines alter the development of people and institutions.
The Capability-Conversion Problem
Measure what develops, not only what performs.
Under what conditions does an increase in machine capability produce a durable increase in human or institutional capability?
Machine capability What an intelligent system can do.
AI-assisted performance What a person or institution can accomplish while using it.
Durable capability What remains as a capacity to understand, judge, create, coordinate, learn, decide, and act.
These are distinct constructs. Machine capability concerns what a system can do. AI-assisted performance concerns what a person or institution can accomplish while using it. Durable capability concerns what remains as a capacity to understand, judge, create, coordinate, learn, decide, and act.
The research question is therefore larger than whether AI improves an outcome. It asks what changes in the person or institution that produced it.
Research domains
Four pillars, multiple levels of analysis
Institutional Integrity
Authority, accountability, organizational design, expertise, decision rights, and institutional learning.
Knowledge Stewardship
Informational abundance, verification, expertise, knowledge creation, preservation, participation, and understanding.
Governed Intelligence
Mandates, bounded authority, evidence of execution, correction, adaptive governance, and the feasible domain of delegation.
Current inquiries
Questions that make capability observable
Capability Conversion
When does AI-assisted performance produce transferable human capability, and when does the benefit remain dependent on the machine?
Expertise and Judgment
How does repeated AI use change expertise formation, independent reasoning, problem formulation, and judgment over time?
Institutions and Delegation
Which institutional arrangements convert machine capability into durable organizational capability while preserving authority and accountability?
Abundance and Scarcity
As information, synthesis, and execution become more abundant, which human capabilities become more consequential?
Knowledge and Understanding
When does abundant machine synthesis deepen expertise and discovery, and when does it weaken the knowledge on which independent judgment depends?
Governed Intelligence
When does greater ability to direct, observe, and revise machine execution expand the work institutions can responsibly delegate?
Research approach
Study what develops.
AI Humanism is interdisciplinary because its questions cross levels of analysis. Method follows the question, but the research program shares several commitments.
Measure capability, not only performance.
Immediate task outcomes and durable human or institutional development answer different questions. Research should distinguish them.
Study development over time.
Expertise, judgment, dependency, institutional memory, and adaptive capacity emerge through repeated interaction rather than a single task.
Examine multiple levels.
Effects on individuals, teams, organizations, professions, and institutions may differ and interact.
Preserve contestability.
Assumptions should remain visible, evidence distinguishable from interpretation, and conclusions open to revision when stronger evidence emerges.
AI Humanism advances when better evidence changes its conclusions.