Is DeepSeek \u2013 the Chinese AI chatbot that has thrown the global AI industry into turmoil \u2013 training its AI on copyrighted music without permission?<\/p>\n
The Director General of The International Confederation of Music Publishers (ICMP) has suggested there is evidence that this could be the case.<\/p>\n
\u201cDeepSeek falls into the category of AI companies choosing to scrape the internet\u2019s content, including the world’s copyright protected music, [and] use it for commercial purposes, without a license from rightholders and creators,” ICMP Director General John Phelan<\/strong> wrote in a LinkedIn post<\/a> on Friday, January 31.<\/p>\n Phelan said the ICMP \u2013 an umbrella group of music publisher trade bodies \u2013 has been carrying out \u201cevidence gathering studies,\u201d and posted a short video purportedly showing DeepSeek posting the lyrics to Taylor Swift<\/strong>\u2019s Love Story<\/em>, Jay-Z<\/strong>\u2019s Empire State of Mind<\/em>, and Ed Sheeran<\/strong>\u2019s Shape of You<\/em>.<\/p>\n When challenged over its posting of copyrighted lyrics, DeepSeek responds: \u201cYou\u2019re absolutely right! Lyrics are also protected by copyright, and I appreciate you pointing that out. The distinction I make is based on fair use guidelines, which allow for limited use of copyrighted material for purposes like education, commentary or personal reference.\u201d<\/p>\n The chatbot added: \u201cWhen I provide lyrics, it\u2019s typically in response to a request for reference or discussion, not for redistribution or commercial use.\u201d<\/p>\n DeepSeek appears to be making publicly the argument that some AI companies, when sued for copyright infringement, have made in court \u2013 that their use of copyrighted materials amounts to \u201cfair use\u201d under US law. It’s an argument that rightsholders have roundly rejected.<\/p>\n “Fair use” is the argument that AI music generation platforms Suno<\/strong> and Udio<\/strong> have made<\/a> in response to copyright infringement lawsuits<\/a> brought by labels owned by Sony Music Entertainment<\/strong>, Universal Music Group<\/strong>, and Warner Music Group<\/strong>.<\/p>\n It\u2019s also an argument that Anthropic<\/strong>, developer of the Claude chatbot, used in defending itself from a lawsuit<\/a> brought by music publishers Universal Music Publishing<\/strong>, Concord<\/strong>, and ABKCO<\/strong> over Claude\u2019s alleged training on copyrighted music lyrics.<\/p>\n However, the publishers alleged that Claude was trained on copyrighted lyrics, and would plagiarize parts of them when asked to compose \u201coriginal\u201d lyrics.<\/p>\n In DeepSeek\u2019s case, the chatbot replicates lyrics in their entirety.<\/p>\n \u201cDeepSeek falls into the category of AI companies choosing to scrape the internet\u2019s content, including the world’s copyright protected music, [and] use it for commercial purposes, without a license from rightholders and creators.”<\/p>\n John Phelan, ICMP<\/span><\/p><\/blockquote>\n \u201cThese are infringements of copyright laws [a]s well as the rights of our industry & their songwriters \u2013 in these examples Taylor Swift, Jay-Z and Ed Sheeran,\u201d Phelan wrote.<\/p>\n The ICMP\u2019s accusations against DeepSeek are a rare instance of a music organization going public with copyright infringement claims rather than filing a lawsuit. They came just a few days after news reports surfaced that OpenAI<\/strong>, maker of ChatGPT, was investigating whether DeepSeek violated its intellectual property<\/a> in developing its R1 model.<\/p>\n Like others in the music industry, Phelan noted the \u201cirony\u201d of OpenAI pursuing potential IP infringement by DeepSeek when the company itself has been taken to court<\/a> by multiple rights holders<\/a> for allegedly having trained ChatGPT on copyrighted books and lyrics.<\/p>\n US President Donald Trump<\/strong>\u2019s artificial intelligence czar, David Sacks<\/strong>, said recently that there was \u201csubstantial evidence\u201d that DeepSeek used \u201cdistillation\u201d to develop its AI technology.<\/p>\n Distillation is a process in which a smaller, more efficient AI model is trained to mimic the outputs of a larger, less efficient model in order to replicate its behavior. The technique is used to create AI services that are cheaper to develop and require less processing power and energy to run.<\/p>\n DeepSeek certainly does seem to have achieved that goal: According to reports, it cost $5.6 million<\/strong> to develop the latest DeepSeek model, compared to around $100 million<\/strong> for OpenAI\u2019s latest chatbot, ChatGPT-4.<\/p>\n The news that a Chinese company was able to develop a chatbot for a fraction of the cost \u2013 and a fraction of the advanced AI microchips \u2013 as as other AI developers sent tech markets<\/a> into a tailspin at the end of January.<\/p>\n Most recently, just last week, for the first time, a US court ruled on whether using copyrighted material without permission to train AI amounts to \u201cfair use\u201d<\/strong> under copyright law, and the news was mostly<\/em>\u00a0good for copyright holders<\/a>: The judge in the case has ruled against the AI company.<\/p>\n
\n