# Bias Audit Documentation

Effective June 11, 2026 | Version 1.0

This document describes how The Ladder AI reviews bias risk in training, assessment, validation, and credential evidence.

## Scope

Bias review applies to:

- Placement conversations.
- Guided training prompts.
- Assessment questions.
- Validation summaries.
- Standards and employment-readiness mapping.
- Credential and transcript language.

## Current Review Model

The Ladder AI uses a combination of internal review, automated checks, user reports, and standards-based validation to identify problems such as:

- Stereotyping or exclusionary examples.
- Unequal treatment across learner backgrounds.
- Overreliance on culturally narrow workplace assumptions.
- Inaccessible or unnecessarily complex language.
- Assessment prompts that reward style over substance.
- Validation summaries that overstate certainty.

## AI and Human Oversight

AI may help generate and review training evidence, but credential records should remain inspectable. Users and institutions should be able to review the transcript, criteria, validation result, and rationale.

## Reporting Concerns

Learners, educators, employers, and procurement teams may report content, accessibility, fairness, or validation concerns through the site issue-reporting process.

## Audit Cycle

The Ladder AI should review bias and fairness risks at least annually and whenever a significant assessment, credentialing, model-provider, or institutional deployment change occurs.

## Current Status

This is an interim internal policy document adapted for The Ladder AI. A formal independent audit or VPAT-style external package should be prepared before large institutional deployments that require it.