# Transparency starts with bounded claims.

This disclosure framework is an editorial draft. It does not assert certifications, completed audits or independently verified assurance.

## AI usage and human review

A publishable disclosure should name the AI-supported activities, the decisions reserved for humans, review responsibilities and the limits of automation. Specific operational claims require approved evidence.

## Data protection and confidentiality

Disclose approved collection purposes, access boundaries, retention rules and authorized processors. Confidential customer evidence and Council discussions must not become public research inputs without authorization.

## Independence and conflicts

Research, membership, sponsorship, subscriptions and independent evaluations serve distinct roles. Commercial relationships must not determine evaluation conclusions. Material conflicts require disclosure and review.

## Corrections and appeals

A challenge should identify the affected claim or report, supporting evidence and requested correction. Preserve the original version, review decision and correction history. Contact Alpha through the existing contact channel.

## Limitations

Public disclosures cannot establish the effectiveness of internal controls. Coverage gaps, unverifiable information and stale evidence must remain visible. No certification claim is made on this page.

## Explore

- [Responsible AI requires accountable governance.](/responsible-ai)

- [Connect principles to evidence.](/responsible-ai/framework)

- [Choose the assessment that fits the evidence.](/responsible-ai/assessment)

- [Ask for decisions, owners and evidence.](/responsible-ai/board-oversight)

- [Transparency starts with bounded claims.](/responsible-ai/transparency)