Generative AI systems, capable of rapidly producing images, video, text, and audio based on textual prompts, have become increasingly prevalent. The success of these systems is indebted to the vast and diverse training data derived from the creative output of millions of creators globally.

Fairly Trained seeks to address a growing divide in the generative AI landscape, distinguishing between companies that obtain consent from training data providers and those that don’t, citing a lack of legal obligation. The organization acknowledges a preference among consumers and companies to collaborate with AI firms respecting creators’ rights and aims to provide clarity in this regard.

The inaugural certification, known as the Licensed Model certification, is awarded to generative AI models that exclusively utilize training data for which the necessary consent has been obtained. This includes various types of licenses, excluding those relying on ‘fair use’ copyright exceptions, signifying a lack of rights-holder consent. Notable among the first nine companies certified by Fairly Trained are Beatoven.AI, Boomy, BRIA AI, Endel, LifeScore, Rightsify, Somms.ai, Soundful, and Tuney, spanning image, music, and singing voice generation.

Fairly Trained’s initiative has garnered support from industry figures and organizations, including the Association of American Publishers, the Association of Independent Music Publishers, Concord, Pro Sound Effects, and Universal Music Group. The certification is seen as a crucial step in fostering transparency and ethical standards in the generative AI sector.

While the first certification focuses on licensing, Fairly Trained plans to delve into more nuanced aspects of dataset acquisition in future certifications, acknowledging the evolving landscape of generative AI and its impact on creators. Companies interested in certification are encouraged to apply, and Fairly Trained aims to provide a valuable badge for consumers and businesses aligned with a commitment to creators’ rights in the generative AI space.

Read more on Fairly Trained through their website here.