13,500+ Artists Sign a Petition to Speak Against AI Data Scraping
- Over 13,500+ artists have come together to petition against AI companies using copyrighted data to train their AI models without a license.
- The petition was initiated by Ed Newton-Rex, a British composer who was also the former audio head at Stability AI.
- After he realized that Stability AI was using copyrighted items without consent to train its AI, he left the company and started on this journey to protect the interests of fellow artists.
More than 13,500 artists and creative professionals have signed a statement against AI data scraping, in an attempt to put an end to this practice.
The unlicensed use of creative works for training generative AI is a major, unjust threat to the livelihoods of the people behind those works, and must not be permitted' - Artist's statement
Popular artists who shared their signature include author Kate Mosse, Kazuo Ishiguro, and Ann Patchett, ABBA's Bjorn Ulvaeus, actor Julianne Moore, and Robert Smith, Radiohead's Thom Yorke and composer/conductor Sir John Rutte.
Organizations such as the European Writers' Council, SAG-AFTRA, Universal Music Group, and the American Federation of Musicians have also joined the petition.
The petition was initiated by British composer Ed Newton-Rex. Interestingly, he was also the head of audio at Stability AI but he resigned last year after the company justified using copyrighted data to train its AI models without license as fair use".
Fair use" is a clause under the US copyright law that permits the use of certain parts of copyrighted items without a license. However, this clause is only applicable in a limited number of cases.And since Stability AI was using those data for financial gains, it clearly doesn't fall under this special exception.
An artist himself, Newton-Rex couldn't put up with this practice and decided to speak up. Along with initiating this petition, he has also started his own venture, Fairly Trained, a company that certifies other companies who train their AI models after obtaining consent from the original creator of the training set.
Ongoing Legal Dispute & InitiativesThis isn't the first time that this issue has been raised. There are already ongoing legal battles between artists and OpenAI, the latter being accused of copyright violations.
Similarly, another petition initiated by the Artist Rights Alliance was signed earlier this year by more than 200 artists. Many popular names such as Nicki Minaj, Billie Eilish, R.E.M, and Stevie Wonder joined the fight to protest against tech companies creating AI music tools.Tools like these are directly competing with the creators, taking away from their livelihood by scraping away content from their own work.
In an industry that's already competitive and new artists are struggling to get a break, there should be no place for a competitor that steals others' work.
In contrast to the popular consensus that AI music tools should be done away with, some artists have switched to the other side and signed agreements with them, perhaps to pay their bills. However, the majority is still against having AI in the creative industry.
But there's a much bigger problem than that - it might already be too late. Take the music industry for example. There are some big AI companies that have already scraped all the music they needed to train their models. Their job is done.So even if the companies are now forced to change their ways for future training, they can simply switch to synthetic data" i.e. using music created by AI models to train more AI models.
To counter this, Newton-Rex believes that synthetic data should be treated the same as human-generated data. After all, that's where it's sourced from.
The post 13,500+ Artists Sign a Petition to Speak Against AI Data Scraping appeared first on Techreport.