The organization will certify artificial intelligence systems that respect copyrights

Generative artificial intelligence models that respect the copyrights of content creators used for educational purposes have begun to attract attention. Certification seal in the United States. The nonprofit organization Fairly Trained is one of the pioneers of this work.

The aim, as the initiative explained this Wednesday (17), is to separate productive AI companies that demand authorization to use protected data from those that do not. In that case, American law does not require developers to disclose the source of contentbut proposals suggesting changes continue.

The first seal issued by the organization Licensed Model certificateIt is given to companies that do not use protected data without permission from the owners. Licenses can be exclusive, expressly permissive, and other types, but will not be granted to developers who claim “fair use” for their work.

At this first moment, nine productive AI companies have already received the Highly Educated sealSuch as Beatoven.AI, BRIA AI, LifeScore and Tuney. These are startups that offer image creation tools using textual description, music, and sound.

Copyright infringement lawsuits

Despite There is currently no legal requirement in the US to obtain consent for data use in generative AI systems., the unbridled use of this information did not please the creators. As a result, the number of lawsuits filed against companies in the sector is increasing.

One of the most notable cases involves OpenAI, the startup behind the popular ChatGPT. New York Times. Newspaper claims developer used unauthorized own materials to train chatbots.

The publication asks about the action involving Microsoft, one of ChatGPT’s main investors: immediate suspension of the use of your content in the AI ​​system. The developer defends itself by saying that it is not possible to train the mechanism without using content made by humans, and claims that copyright laws are respected.

Fairly Trained was founded by former Stability AI VP Ed Newton-Rex. After leaving his old job, the manager They exploited creators and ended up competing with their work.