Among AI music generators capable of creating entire songs with just a prompt, Suno and Udio are considered by many to be among the best.<\/p>\n
They are also considered by many to have used copyrighted music to train their AI models, without authorization.<\/p>\n
Now, the Recording Industry Association of America<\/strong> (RIAA<\/strong>) is coordinating lawsuits on behalf of major recording companies against the two AI companies.<\/p>\n Sony Music Entertainment<\/strong>, Universal Music Group<\/strong>\u2019s UMG Recordings,<\/strong> and Warner Records Inc.<\/strong> are among the plaintiffs in two new federal copyright infringement lawsuits filed on Monday (June 24).<\/p>\n The lawsuit against Suno<\/strong> was filed in the US District Court for the District of Massachusetts, while the lawsuit against Udio<\/strong> was filed in the US District Court for the Southern District of New York, over what the RIAA calls \u201cmass infringement of copyrighted sound recordings copied and exploited without permission by two multi-million-dollar music generation services.\u201d<\/p>\n \u201cAI companies, like all other enterprises, must abide by the laws that protect human creativity and ingenuity,\u201d the complaints against Suno and Udio state. \u201cThere is nothing that exempts AI technology from copyright law or that excuses AI companies from playing by the rules.\u201d<\/p>\n That assertion is likely to be the key point of contention in the lawsuits. US courts have not yet ruled that using copyrighted materials to train AI amounts to copyright infringement.<\/p>\n A number of tech companies involved in developing AI tech, including Anthropic<\/strong><\/a> and Google<\/strong><\/a>, have argued that using copyrighted materials to train AI should be considered a \u201cfair use\u201d exemption to copyright law, as, in some circumstances, it\u2019s permissible to use copyrighted materials in the creation of new technologies or products.<\/p>\n However, music companies and other rights holders argue that this doesn\u2019t apply to generative AI, as this technology can be used to create material that competes directly with the copyrighted material being used. AI music generators are capable of creating songs that compete in the marketplace against the copyrighted songs used to train them.<\/p>\n \u201cWhen those who develop such [AI music services] steal copyrighted sound recordings, the [services\u2019] synthetic musical outputs could saturate the market with machine-generated content that will directly compete with, cheapen, and ultimately drown out the genuine sound recordings on which the [services were] built,\u201d the legal complaints state.<\/p>\n \u201cThe music community has embraced AI and we are already partnering and collaborating with responsible developers to build sustainable AI tools centered on human creativity that put artists and songwriters in charge.”<\/p>\n Mitch Glazier, RIAA<\/span><\/p><\/blockquote>\n In a statement shared with media on Monday, the RIAA asserted that the use of copyrighted songs to train AI music generators fails all four of the tests used by US courts to determine whether the \u201cfair use\u201d exemption applies.<\/p>\n \u201cFair use is not available when the output seeks to \u2018substitute\u2019 for the work copied. And Suno and Udio have, in their own words, conceded that is exactly what they intend,\u201d the RIAA asserted.<\/p>\n \u201cThe music community has embraced AI and we are already partnering and collaborating with responsible developers to build sustainable AI tools centered on human creativity that put artists and songwriters in charge,\u201d RIAA Chairman and CEO Mitch Glazier<\/strong> said in a statement.<\/p>\n \u201cBut we can only succeed if developers are willing to work together with us. Unlicensed services like Suno and Udio that claim it\u2019s \u2018fair\u2019 to copy an artist\u2019s life\u2019s work and exploit it for their own profit without consent or pay set back the promise of genuinely innovative AI for us all.\u201d<\/p>\n The recording industry trade group also claimed that both Suno and Udio have \u201cadmitted\u201d to using copyrighted material to train their AI.<\/p>\n The RIAA cited comments by an early investor in Suno, Antonio Rodriguez<\/strong>, who all but admitted<\/a> that Suno had used copyrighted materials when he stated that he opposed Suno signing licensing deals with music copyright owners, saying \u201cif [Suno] had deals with labels when this company got started, I probably wouldn\u2019t have invested in it. I think they needed to make this product without the constraints.\u201d<\/strong><\/p>\n The group also asserted that comments by Udio executives that their AI had been trained on \u201ca large amount of publicly-available and high-quality music\u201d that was \u201cobtained from the internet\u201d and was the \u201cbest quality music that\u2019s out there\u201d is essentially an admission that they used \u201cthe copyrighted sound recordings of the RIAA\u2019s member companies.\u201d<\/p>\n The RIAA further asserted that ($125 million-backed<\/a>) Suno<\/strong> and (Andreessen Horowitz-backed<\/a>) Udio<\/strong> had been \u201ccaught\u201d using copyrighted materials.<\/p>\n \u201cThey left producer tags on their output. \u2018CashMoneyAP\u2019 and \u2018Jason Derulo\u2019 tags are present in Suno\u2019s outputs,\u201d The RIAA stated.<\/p>\n \u201cUsers have generated sound-a-likes of numerous sound recordings using Suno and Udio, including The Temptations\u2019 My Girl<\/em>, Green Day\u2019s American Idiot<\/em>, Mariah Carey\u2019s All I Want for Christmas<\/em>, along with recordings by Chuck Berry, James Brown, and others.\u201d<\/p>\n \u201cThese lawsuits are necessary to reinforce the most basic rules of the road for the responsible, ethical, and lawful development of generative AI systems and to bring Suno\u2019s and Udio\u2019s blatant infringement to an end.\u201d<\/p>\n Ken Doroshow, RIAA<\/span><\/p><\/blockquote>\n This echoes the forensic analyses of Suno- and Udio-generated music carried out by Ed Newton-Rex<\/strong>, the former head of music at Stability AI<\/strong>, who founded<\/a> the non-profit ethical AI accreditation organization Fairly Trained<\/strong>.<\/p>\n In an analysis<\/a> published by MBW, Newton-Rex found striking similarities between music created by Suno and the works of Ed Sheeran, ABBA, Oasis and others. In a separate analysis<\/a>, he found similarities between Udio-generated music and the works of John Lennon, Natalie Imbruglia, Coldplay and others.<\/p>\n \u201cIn addition, in some instances, the defendants\u2019 services produce vocals that are indistinguishable from famous recording artists, including Lin-Manuel Miranda, Bruce Springsteen, Michael Jackson, and ABBA,\u201d the RIAA asserted.<\/p>\n \u201cThese are straightforward cases of copyright infringement involving unlicensed copying of sound recordings on a massive scale. Suno and Udio are attempting to hide the full scope of their infringement rather than putting their services on a sound and lawful footing,\u201d said Ken Doroshow<\/strong>, the RIAA\u2019s chief legal officer.<\/p>\n \u201cThese lawsuits are necessary to reinforce the most basic rules of the road for the responsible, ethical, and lawful development of generative AI systems and to bring Suno\u2019s and Udio\u2019s blatant infringement to an end.\u201d<\/p>\n The RIAA-led lawsuits against Suno and Udio come after numerous of other lawsuits were launched in the US against AI developers over the past year.<\/p>\n Among them is a lawsuit that includes Universal Music Group<\/strong>, Concord<\/strong> and ABKCO<\/strong> as plaintiffs, arguing<\/a> that Anthropic AI\u2019s chatbot Claude has been ripping off copyrighted lyrics and passing them off as original.<\/p>\n
\n