AI & Data Disclosure

Current website

This public website does not run an AI model on visitor input and does not accept commitment text. No production AI provider, model release, processing region, or retention posture has been activated for public users.

Planned first-release role

For a commitment a user chooses to submit, a bounded language model may propose:

• what appears to be requested and source text supporting that proposal;
• timing, visible effort, hidden-work candidates, and consequential ambiguity;
• an explanation of results already calculated by deterministic code;
• an editable response draft constrained by the user’s recorded choice;
• a summary of already-canonical weekly facts.

What AI will not decide

• what a part of the user’s life is worth;
• which protected priorities the user should sacrifice;
• whether an unsupported inference becomes a confirmed fact;
• the canonical capacity, planning range, displacement, entitlement, or workflow state;
• whether anything is sent, scheduled, purchased, or changed externally.

Human and deterministic review

Models propose. Consequential facts require evidence or user confirmation. Deterministic, versioned code owns the numeric Decision Packet and canonical state. The user chooses the response. In the first release, a response draft remains unsent unless the user sends it personally.

Uncertainty and error

AI output can be incomplete or wrong. “Not enough information” is a valid result. A production disclosure will identify how users inspect evidence, correct information, report a problem, export data, and request deletion.

Provider and data controls

Before AI processing opens, this page will name the approved provider, material service category, model release policy, processing region, training posture, abuse-monitoring and application-state retention, subprocessors, and deletion behavior. No Zero Data Retention claim will be made unless the actual account and every used capability support it.

Longer-term direction

Governed AI action and verified time return are not part of the first release. The intended future system would show exact proposed work, expected cost, potential time return, authority, read-back, rework, and whether return is potential, estimated, verified, disputed, or unverifiable.