Working Paper Series
The ALI Working Paper Series publishes in-progress research on AI alignment governance to invite peer critique, domain expert engagement, and co-authorship. Papers are shared as clearly labeled working drafts. The work is longitudinal and multi-vertical — each paper a building block in a shared research program. If the framework interests you, the co-author slot is open for a reason.
What We Mean When We Say Alignment: Definitional Fragmentation as a First-Order AI Governance Failure
The dominant communities engaged with AI alignment are using the same term to describe properties so different in kind that they resist comparison — a condition we call definitional fragmentation, and a first-order coordination failure that structurally impedes governance before any technical question is reached. This paper proposes a four-part operational taxonomy distinguishing technical, behavioral, governance, and societal alignment, with empirical grounding and falsification conditions for each, mapped against the structural requirements that adequate governance would need to satisfy.
Each paper in the series carries an open co-author slot in a specific domain. Paper 1 is seeking a researcher with depth in governance, regulatory frameworks, or science and technology studies — someone who can stress-test the four-part taxonomy against institutional and political theory, and contribute to the multi-vertical research program this paper initiates.
This draft is shared openly to invite substantive engagement — critique, correction, or a proposal to build on the framework together. Reach out directly with your background and what you’d bring to the work.
- Paper 2 Inaugural vertical case study · in development
- Paper 3 Financial services and algorithmic risk · planned
- Paper 4 Healthcare and clinical decision support · planned
- Paper 5 Critical infrastructure and autonomous operations · planned
- Paper 6 Legal systems and adjudicatory AI · planned