What it is
Fairly Trained is a certification program for models and model builders that claim to use properly licensed or consented training data. It belongs in this catalog because it turns sourcing norms into a concrete third-party signal that partners, customers, and policymakers can point to, even if certification counts still understate the full set of participating models.
Watchouts
Published announcement counts may understate the current roster; Fairly Trained's current certified-models page appears larger, but it is undated, so this metric stays conservative.
Evidence trail
- Jul 31, 2024 Individual model certification updated
- Mar 19, 2024 Five more companies received certification
- Jan 16, 2024 Certification launched for generative AI companies
Adoption signals
Users
16+ announced certified entities
Derived from nine launch certifications, five additional certified entities announced in March 2024, and two more individual-model certifications announced in August 2024.
- Individual model certification updated Jul 31, 2024
- Five more companies received certification Mar 19, 2024
- Certification launched for generative AI companies Jan 16, 2024