Disclosure Threshold
The point in a customer interaction, determined by the consequentiality of the decision rather than the channel it occurs on, at which failing to disclose autonomous operation becomes a trust liability rather than a neutral design choice.
The Disclosure Threshold rejects both a blanket disclose-everything rule and a blanket disclose-nothing default, replacing either with a specific, testable standard: consequentiality, not channel or convenience, determines whether concealment is a neutral omission or a trust violation. A customer whose password was reset correctly has received exactly what they wanted regardless of who performed the reset; a customer disputing a charge is implicitly trusting that a reasonable, accountable process produced the outcome, and concealing that no human weighed their specific circumstances withholds information they would reasonably want.
The risk asymmetry is the strongest argument for a fixed position rather than a live judgment call. Low-consequence concealment discovered later costs little. High-consequence concealment discovered later compounds specifically because the stakes were real — the customer does not just learn they spoke to an agent, they learn the business let them believe a person had reviewed their case at exactly the moment that belief mattered most. This connects to the same discipline Redundancy Dividend applies to untested reliability claims: a belief discovered false at the worst possible moment is worse than no belief at all.
The Disclosure Threshold shares its classification work with Deterministic Outcome designation: a decision class verified as producing a Deterministic Outcome typically sits below the threshold because the customer can verify the result themselves; a decision class requiring genuine judgment typically sits above it. This makes the threshold a specification point at Full-System Design time rather than a case-by-case decision, and it becomes the evidentiary link the Accountability Trace draws on when a decision above the threshold later requires a demonstration that disclosure actually occurred.
Application
Interactions with a Deterministic Outcome — a password reset, a delivery status check — sit below the threshold, since the customer's concern is the verifiable result rather than who produced it. Interactions involving judgment, empathy, or consequential decisions the customer cannot verify themselves — a disputed charge, a denied claim — sit above the threshold and require disclosure, because concealment withholds information the customer would reasonably want to assess fairness.
Context
The Disclosure Threshold takes an explicit position the rest of the corpus had left open: whether a customer needs to know no human is involved. The risk is asymmetric — concealment discovered at a low-stakes interaction is minor, but concealment discovered at a consequential one compounds, because the customer learns the business let them believe a person had weighed their specific situation at exactly the moment that belief mattered most. The practical test: if the customer would feel reasonably misled upon learning the truth, disclose; if they would reasonably shrug, disclosure is optional.
This term is machine-readable
Any MCP-compatible AI assistant can retrieve the canonical definition of Disclosure Threshold at inference time — no training approximation.
Related Terms
In the Log
First used: July 2026
Edition 1 · updated July 2026