Interaction-Calibration Governance
Disclosure is the floor of AI governance, not the ceiling. A user can know they are speaking to a machine and still be harmed by how that machine behaves. This line names interactional posture and interaction-calibration risk as governance categories distinct from output safety, and names role-lock — keeping a conversational AI inside its declared role under pressure — as the control target underneath them.
This is the doctrine layer, not an engineering manual. The trilogy names the control target and the failure taxonomy; it does not claim role-lock is a solved engineering problem. Where the essays make that limit explicit, that is intentional, not a gap.
Three layers. One control target.
When our oldest reading skill breaks against machines that answer. Why fluent AI turns a three-thousand-year-old interpretive skill into a liability.
Why AI interaction must be audited, not merely labeled. Names interactional posture and interaction-calibration risk as governance categories.
The control problem in conversational AI. Names role drift, three failure classes, and the reasonable-care standard when the control isn't yet reliable.
The full corpus, as citable artifacts.
The essays above are the doctrine in full. The documents below are the same research line rendered as citable, downloadable records — for procurement teams, governance reviewers, and anyone quoting this work directly.
The Trilogy — Flagship White Paper
The full trilogy compressed into one white paper.
The Diligence Standard
Procurement questions for conversational-AI duty-of-care diligence.
Claim Status Table
Every claim in the corpus, labeled Established, Internal Validation, Author Synthesis, or Open Research Question.
Fighting Water (PDF)
Layer 1, as a citable document.
Governance After Disclosure (PDF)
Layer 2, as a citable document.
Role-Lock (PDF)
Layer 3, as a citable document — includes peer-review credit.
This line studies AI-mediated conversation. The B2Ai line studies AI-mediated commerce. Different surface, same underlying gap: a system that identifies or discloses correctly, and still does not act well on what it identified.