Can Semantic Geometry Teach an AI Judgement?

Abstract

How can an AI agent determine what rules to follow? One rule permits an action. Another imposes a condition, exception, or conflicting obligation. Deterministic systems can resolve those relationships when they have been specified. When they remain implicit in language, an agent can follow one rule while missing another that should stop it. Refusing every unresolved action avoids that risk, but also blocks permissible actions.

We wanted the agent to make the distinction and still act. Our initial hypothesis was that geometric measurements could supply a basis for judgment. We represented actions and policies as vectors, then tested whether their geometry could identify governing policies and interpret the action's relation to them.

Across four studies, the tested approaches did not establish reliable pre-action judgment. In the final synthetic study, a lexical router recovered every governing and blocking policy while reducing median policy checks by 97.7%. The composed pipeline nevertheless escalated all 2,304 test actions, including those it should have allowed. Supplying every policy to the same downstream mechanism changed no decision. Finding the policies had not solved the problem of interpreting them.

This result led us to revise our hypothesis: judgment in AI agents requires developing a consequence graph. Such a graph would connect the actor and authority to policy conditions, exceptions, and the changes an action would produce. Follow-on studies will ask whether making those relationships explicit helps the agent distinguish when to act, stop, or seek review.

Fields and Methods

artificial intelligence, semantic geometry, policy retrieval, pre-action judgment, governed autonomy.

• four bounded studies separating compact representation, component measurements, composed decisions, and routing versus interpretation
• synthetic action-policy cases with frozen calibration and evaluation splits
• comparisons of geometric measurements, lexical routing, and downstream ACT, HOLD, or ESCALATE decisions
• an all-policy comparison to distinguish retrieval failure from downstream interpretation failure

Collaborator Profile

Researchers in AI agents, policy evaluation, semantic representation, and structured reasoning interested in testing how agents interpret conditions, exceptions, authority, and consequences before acting.

Validation Needed

• The tested approaches did not establish reliable pre-action judgment; these bounded studies do not rule out all geometric methods.
• The consequence-graph hypothesis is a proposed follow-on, not a tested or validated result.
• The public companion supplies frozen evidence and bounded offline recalculation inputs; it does not establish full experimental replication.

Publication Posture

This is a public institute preprint, not a peer-reviewed finding. Its permanent arXiv record is arXiv:2610.07249 in cs.AI. The arXiv record is the primary public source.

Citation

Thomson D. Nguy. Can Semantic Geometry Teach an AI Judgement? arXiv:2610.07249 [cs.AI].

Research Materials

The public GitHub companion contains frozen scientific evidence, saved results, and offline recalculation inputs. Version 0.1.0 states its coverage and missing inputs explicitly. Full experimental replication is not claimed.

Research Correspondence

For collaboration inquiries or research correspondence about this work, contact contact@theradiantinstitute.org.