Decision Closure Framework (DCF): Specification and Companion Paper

This record contains the Decision Closure Framework (DCF) specification (Committee Draft v1.3) together with its companion paper.

 

The Decision Closure Framework defines a formal model for continuous decision legitimacy in AI and socio-technical systems. It introduces three core concepts:

 

- Continuous Closure: the requirement that decision legitimacy must persist throughout execution, not only at the point of decision  

- Authority Rebinding: the explicit reassignment of decision authority when legitimacy conditions fail during runtime  

- Failure Transition Control (FTC): a governance layer that regulates system behavior during transitions from legitimate to illegitimate execution states  

 

The framework addresses a critical gap in current AI governance approaches. Existing systems focus on decision correctness, constraint enforcement, and post-hoc accountability, but do not formally govern what happens when the conditions underpinning a decision change during execution.

 

DCF shifts the focus from:

"How do we ensure decisions are correct?"

to:

"What determines whether a decision is allowed to be executed at all—and who has authority when that condition changes?"

 

The included Committee Draft provides a normative specification aligned with emerging standards in AI governance (including ISO/IEC 42001 and the EU AI Act), while the companion paper offers conceptual grounding, problem framing, and interpretative context.

 

This work is part of a broader governance architecture including:

- system-level capability certification  

- decision-level identity and legitimacy protocols  

- observation-level drift and integrity diagnostics  

 

Status: Committee Draft (CD). This document is released for public review, discussion, and iterative refinement. It must not be cited as a finalized standard.

 

Keywords: AI governance, decision integrity, authority models, execution control, failure handling, continuous closure, AI safety architecture