Updating our individual model certification

Today we’re announcing an update to our L (Licensed Model) certification - specifically, the way generative AI companies can apply to certify individual AI models.

AI companies often use architectures that include multiple models. The different models play different roles, and are often trained on different data. So an image generation architecture, for example, might use three different models: a U-Net, a text encoder and an autoencoder. Each model has a different function. Every AI company uses a different architecture, often including third-party models, meaning the nuances of training data sourcing can be complex.

Companies could already get individual models certified by Fairly Trained. From today, companies awarded individual model certification will receive a new certification badge, and will be required to be transparent with users about which elements of their architecture are and are not certified. We think this transparency is critical to let users make informed decisions about which AI products to use.

We are pleased to announce that two companies have received this certification.

Generative music platform Jen has had their omnidirectional audio diffusion models certified by Fairly Trained. Shara Senderoff, Co-Founder and CEO of Jen, says: "It's been our top priority at Jen to architect an approach that leads the way for ethically-trained models to establish a dominant presence at the forefront of their industries. We believe that this prioritization not only adheres to responsible practices but also positions models to be trusted and widely adopted. We're incredibly proud to have our audio models certified by Fairly Trained, a meaningful alignment in our commitment to transparency."

And AI audio tool company Kits AI has had its voice and instrument models certified, including its verified voice library, which compensates artists who trained their voices on the Kits platform per-download. Pippin Bongiovanni, Kits' Director of Marketing, says: "At Kits, a big part of our mission is to empower vocalists and creators with AI tools that respect artistry, fairly compensate artists, and promote a sustainable future for music—something we believe Fairly Trained certifications move us all closer to."

We hope this update to our certification will help highlight more AI companies that are making significant efforts to treat creators fairly when sourcing training data.

We are also pleased to announce three new supporters of Fairly Trained: NAVA, the Society of Composers & Lyricists, and the Artists Rights Society.