Components › Transparency & Disclosure › Reasoning Disclosure
TD07 · Transparency

Reasoning Disclosure

Purpose A collapsible panel that surfaces how the AI moved from input to output — the considerations weighed, the alternatives discarded, and the trade-offs accepted.
Use when AI reorders, ranks, filters, pre-selects, or frames choices based on user history, profile, or predicted preferences — at any decision point where the AI is influencing what options the user sees or how prominently they appear.
Do not use For AI outputs that are purely informational without affecting the structure of available choices.
Failure mode prevented Users make decisions they believe are self-directed, unaware the AI has structured the choice environment to favor certain outcomes — eroding autonomy without consent or accountability.
Example Dark Academia
Dark Academia
The Atelier
Neon Tide
The Archive
Candlelight
The Conservatory
For your enterprise context, I'd recommend React with TypeScript. It fits your team's existing Java background and gives you the strongest long-term maintainability story at scale.
Overview
Design rule

Reasoning disclosure is a trust instrument, not a transparency checkbox. Show the reasoning that helps users calibrate whether to act on the output — not everything the model generated. When in doubt, default to collapsed.

Theming API

Override these CSS custom properties to apply your own visual treatment without touching the component's behavioral contracts.

Shared semantic tokens

Token Controls
--rad-border-subtle Panel border and node dividers
--rad-radius-card Panel border radius
--rad-color-text-primary Conclusion node text
--rad-color-text-secondary Default node text
--rad-color-text-meta Toggle label, discarded node text
--rad-color-approve Node labels (Context, Considered, Trade-off)
--rad-color-emphasis Conclusion node label accent
--rad-color-focus Toggle focus outline

"*Reasoning disclosure is a trust instrument, not a transparency checkbox. Show the reasoning that helps users calibrate whether to act on the output — not everything the model generated.*"

← Previous Operational Transparency Next → Human-in-the-Loop

Intellectual Property

RAD is the original work of Jackie Curry. All rights reserved. No portion may be reproduced, adapted, or incorporated into any product or system without express written permission.

Permitted: citation in academic or editorial contexts with full attribution.

Licensing

© 2025 Jackie Curry. All rights reserved. Publication date: 2025.

For licensing inquiries, connect on LinkedIn →