Distributed Agency in Human-AI Systems: A Framework for Analyzing Authorship, Control, and Autonomy
On the Calculation of The Human-AI Agency Spectrum A framework for measuring authorship distribution in AI-participated creative work.
Distributed Agency in Human-AI Systems: A Framework for Analyzing Authorship, Control, and Autonomy
By Mark Ghuneim | Narrative.new | 2026 | License: CC BY 4.0
What This Is: A six-layer model (L0-L5) that maps qualitative standards like “human-led” and “significant human control” to operational configurations. It synthesizes:
Legal requirements — USCO, EU AI Act Article 50 Professional standards — WGA, SAG-AFTRA, DGA Cognitive research — on human-AI interaction and creative intent Market analysis — compliance constraints that create market tiers
The result is a shared vocabulary for discussing agency that connects to the rules already governing professional work.