Architectural Wisdom: A Framework for Governing Optimization in AI Systems
Edward Y. Chang
Why It Matters
What makes this one worth your time
This work is relevant as it addresses critical failures in AI systems that arise from poorly defined objectives, potentially guiding future designs for more robust and ethical AI.
A framework for integrating wisdom into AI optimization processes.
Summary
The paper proposes a framework called 'architectural wisdom' that aims to govern optimization in AI systems by addressing the structural failures that arise from under-specified objectives.
Key contributions
- Introduction of the architectural wisdom framework for AI systems.
- Identification of three structural commitments: temporal horizon, relational boundary, and irreversibility.
- Development of four components to compute a wisdom tuple for AI decision-making.
Notable insights
- The distinction between wisdom and intelligence as architectural properties could reshape how we design AI systems.
- The framework's emphasis on explicit structural commitments may provide a new avenue for improving AI governance.
Possible limitations
- Not stated in the abstract.
Abstract
arXiv:2606.16319v1 Announce Type: new Abstract: Modern AI systems exhibit structural failures that capability scaling alone does not reliably fix: they optimize under-specified objectives with no architectural mechanism to question whether the objective should be optimized at all. Engagement maximization can amplify harmful pathways; tool-using agents can commit irreversible actions; preference-trained language models can become sycophantic. We argue that this failure is a wisdom problem, not an intelligence problem. We use "wisdom" in a deliberately architectural sense, not as a claim about virtue, consciousness, or moral omniscience. Intelligence accepts a goal and optimizes within it; wisdom interrogates whether the goal should be optimized at all. The two are separable architectural properties. We propose architectural wisdom as a corrigible objective-governance layer above the optimization substrate. The layer makes three structural commitments explicit and nondegenerate before any action: temporal horizon, relational boundary, and irreversibility. It is realized by four components (Structural Utility Transform, Moral Admissibility Interface, Arbitration and Escalation Controller, Value Revision Channel) that compute a six-coordinate wisdom tuple over horizon, relational coverage, irreversibility, admissibility, value revision, and auditability. We motivate the architecture by eight cases drawn from contemporary AI failures, secular wisdom traditions, and hard ethical situations, and defend the distinction against the intelligence-completeness thesis using goal-questioning over goal-taking, Bostrom's orthogonality, structural separation in our exemplar cases, and persistent failure modes despite capability scaling. The framework is the conceptual contract for a larger architecture whose formal specifications and empirical validation are developed in subsequent work.