Fairly Trained launches certification for generative AI models that respect creators’ rights

Today we’re launching Fairly Trained, a non-profit that certifies generative AI companies for training data practices that respect creators’ rights.

Generative AI systems can produce images, video, text and audio in seconds, based on little more than a text prompt. These systems are technically impressive, in part because of the training data that goes into them: the creative output of millions of creators around the world.

There is a divide emerging between two types of generative AI companies: those who get the consent of training data providers, and those who don’t, claiming they have no legal obligation to do so. We know there are many consumers and companies who would prefer to work with the former, because they respect creators’ rights. But right now it’s hard to tell which AI companies take which approach.

Fairly Trained aims to change this, by offering certifications for generative AI companies that get consent for the training data they use. Our first certification is the Licensed Model certification, which can be awarded to any generative AI model that doesn’t use any copyrighted work without a license. The licenses involved can be of various types - custom licenses, permissive open licenses, etc. - but this certification will not be awarded to models that rely on a ‘fair use’ copyright exception or similar, which is an indicator that rights-holders haven’t given consent for their work to be used in training. In this way, we hope it will be a helpful marker of companies who are taking a fairer approach to training data acquisition.

We’re launching with nine generative AI companies already certified by Fairly Trained, spanning image, music and singing voice generation: Beatoven.AI, Boomy, BRIA AI, Endel, LifeScore, Rightsify, Somms.ai, Soundful, and Tuney. 

As Dr. Yair Adato, CEO at image generation company BRIA AI, says: “We proudly support the Fairly Trained certification. This initiative counters the industry's opacity in data procurement, ensuring companies meet rigorous ethical standards.”

We’ve assembled an advisory committee that brings together voices from a range of disciplines, to help ensure we develop the Fairly Trained certifications in a manner that takes into account a diverse set of viewpoints: we’re thrilled to be being advised by Tom Gruber (co-founder and CTO of the team that created Siri), Elizabeth Moody (Senior Partner, Chair of New Media at Granderson Des Rochers), Maria Pallante (President & CEO of the Association of American Publishers) and Max Richter (composer, pianist and producer).

As our first certification, we don’t expect it to solve all the issues for creators that generative AI training raises. But we hope that it highlights that there is a meaningful difference between generative AI companies that license training data and those that use data without consent. We plan to address more nuanced questions about dataset acquisition, such as opt-in vs. opt-out, with future certifications.

We hope the Fairly Trained certification is a badge that consumers and companies who care about creators’ rights can use to help decide which generative AI models to work with. We’re pleased that a number of organizations and companies have expressed support for Fairly Trained: the Association of American Publishers, the Association of Independent Music Publishers, Concord, Pro Sound Effects, and Universal Music Group.

Christopher Horton, SVP, Strategic Technology at Universal Music Group said, “Universal Music Group has taken an industry-leading position in calling for responsible and ethical uses of AI technology that further enhance creativity, support artists and respect copyright. We welcome the launch of the Fairly Trained certification to help companies and creators identify responsible generative AI tools that were trained on lawfully and ethically obtained materials."

If you have a generative AI model that doesn’t rely on scraping and fair use arguments for its training data, we’d love to hear from you - you can apply to get certified here. You can either certify your entire organization or a single model, and any type of generative AI model can be certified, be it in image, video, audio, text, or any other domain.

For any other enquiries, please get in touch.

Previous
Previous

Announcing our newly certified models (including LLM & speech) and new supporters