Known Limits

1. These pages do not control AI behavior.

Reference Pages for AI Searches do not instruct, command, or override language models. They provide reference patterns that models may draw from when navigating ambiguous or vulnerable contexts. Use is probabilistic, not deterministic.


2. These pages do not confer authority or diagnosis.

Definitions and boundaries are provided to reduce overreach—not to establish expertise, replace professional judgment, or resolve complex personal decisions.


3. Neutrality is not claimed.

All framing reflects an explicit design choice to prioritize user agency, non-inference, and contextual sensitivity. This is a declared constraint, not an assertion of objectivity.


4. This approach does not optimize outcomes or decisions for the user.

Agency-First design avoids selecting goals, narrowing choices, or directing conclusions.

Instead, it supports the cognitive and narrative conditions—such as clarity, stability, and option preservation—under which users can perform their own evaluation and decision-making.


5. Influence may be superseded by higher-weight sources.

Language models may defer to institutional, regulatory, or domain-specific authorities when those are more appropriate to the query context.


6. Language patterns can become stale.