TapTechNews August 2nd news, the music AI company Suno released a blog post yesterday (August 1st) in response to the lawsuits filed by the three major record companies, admitting that it had used the record contents of these three major record companies to train the AI, but the company believes that according to the fair use principle, this is legal.
The Recording Industry Association of America (RIAA) sued two music-generation start-ups, Udio and Suno, on June 24th, accusing these two companies of using copyrighted music to train AI models.
These musics involve those from Universal Music Group, Sony Music, and Warner Music, the three record companies.
Suno company responded that in the process of using tens of millions of records to train the AI model, there were indeed record contents from the above three record companies, but according to the fair use principle of the US copyright law, such activities belong to the scope of fair use, which means that no license is required, and accused the record companies of abusing copyright.
TapTechNews translated part of the content of the report submitted by Suno as follows:
In the copyright law, if the protected work is copied for back-end technology and not disclosed to the public and not used to create a new product service that is ultimately not infringing in the service, it belongs to fair use.
Since the US Congress enacted the first US copyright law in 1791, in the 233 years since then, no contrary conclusion has been reached in any case.
The final conclusion is that making intermediate copies of protected works is permitted and not actionable.
Mikey Shulman, the CEO and co-founder of Suno, published a blog post on the same day he submitted legal documents, which continued to say:
We train our models based on the medium and high-quality music that can be found on the open Internet... Much of the content on the open Internet does contain copyrighted materials, and some of those materials are owned by large record companies.
Using the data on the open Internet to train the artificial intelligence model is no different from kids creating their own rock songs after listening to rock music. Learning is not infringement. It wasn't in the past, and it isn't now.