Production AI Institute · PSF v1.1 open standard
AI Right-To-KnowAI Data Use IndexCheck My AI ToolsAgent ReadinessPublic BenchmarkContactGlobal standard · Australia/APAC founded
<- AI Data Use Index
Methodology

How the AI Data Use Index is maintained.

The index records what the public can actually tell from the public record. It grades disclosure clarity, not secret intent.

Review rules

What PAI does and does not claim.

Use official provider sources wherever possible.

Review at the product level, not with vague company-wide claims.

Separate stated policy from PAI commentary.

Record the review date visibly on every entry.

Do not infer hidden behavior from silence.

Update quickly when a provider changes the public record.

Status definitions

Four labels, used narrowly.

Uses data

The provider publicly states some form of training reuse for the product or relevant data category.

User-controlled

The public answer materially depends on a user or admin setting, permission, or privacy mode.

Does not use

The provider publicly states that the reviewed product/data category is not used for model training.

Unclear

The public record does not yet support a simple enough answer at the product level.