AI startup says scraping songs on internet is 'fair use'
Music-focused artificial intelligence (AI) startups, Udio and Suno, are currently embroiled in legal battles with prominent music labels including Universal Music Group, Warner Music Group, and Sony Music Group. The lawsuits were initiated in June this year, alleging that these companies trained their AI models using copyrighted content scraped from the internet. Suno has openly admitted to the court that its AI bots do indeed scrape copyrighted songs for training purposes but justified its actions as 'fair use.'
Suno's AI model trained on millions of copyrighted recordings
In a court filing, Suno acknowledged that its AI model was trained on tens of millions of recordings, some of which are likely owned by the plaintiffs in this case. The company clarified that its training data encompasses virtually all music files of reasonable quality available on the open internet. This dataset potentially includes millions of unauthorized copies of songs. Despite this admission, Suno maintains that its actions are protected under the principle of fair use.
Suno defends its actions as part of 'back-end technological process'
"It is fair use under copyright law to make a copy of a protected work as part of a back-end technological process, invisible to the public, in the service of creating an ultimately non-infringing new product," Suno said. This defense is based on the premise that since AI-generated tracks it produces don't include samples from original songs, obtaining those tracks for AI model training isn't problematic.
Industry body refutes Sono's 'fair use' defense
The Recording Industry Association of America (RIAA), which initiated the lawsuit, has strongly refuted Suno's defense. A spokesperson for the organization labeled the defendants' actions as "evading and misleading," and argued that "their industrial scale infringement does not qualify as 'fair use.'" The industry body also criticized Udio and Suno for failing to obtain consent before using artists' work, stating that many of their competitors have already done so.
