① According to reports, AI companies such as OpenAI, Microsoft, Google, Suno, and Udio have all received warning letters; ② Sony Music Group has also made claims against streaming platforms such as Apple and Spotify; ③ As AI develops in multiple modes, the copyright war in the audio field is heating up rapidly.
Financial Services Association, May 17 (Editor Shi Zhengcheng) According to the latest news on Thursday, Sony Music Group (Sony Music Group), one of the world's largest record companies, is taking action to ensure that its music copyright is not ruthlessly abused by technology companies in the AI torrent.
According to reports, Sony Music has sent letters to more than 700 artificial intelligence companies and streaming platforms around the world, warning them not to use Sony Music's intellectual property rights without clear authorization to train AI models, and demanding that they immediately stop infringement.
Sony Music also published a “Statement on Opting Out of AI Training” on its official website, publicly calling for infringing global developers to immediately stop using infringing information to train AI and reserve the right to trace past infringements. Of course, companies that have been personally sent warning letters by Sony will also have a bigger problem.
(Source: Sony Music Group)
“There is reason to believe that you have violated rights”
According to people familiar with the matter, companies developing AI systems such as OpenAI, Microsoft, Google, Suno, and Udio have already received letters from Sony. At the same time, Sony also sent separate letters to Apple and Spotify requesting that streaming platforms take the best measures to protect artists and songwriters from being illegally used by unauthorized AI companies to train models. Sony has also specifically requested these internet platforms to update their terms of service, clearly stating that crawling and training of their content is not permitted.
Sony's actions have also involved a dark line in the AI industry in recent years — how exactly did the training model data come from? For example, in recent months, many AI companies have launched the “Wensheng Video” app, but when asked about the source of the training materials, the vast majority of companies only looked back and forth at him.
According to a letter read by the media, Sony said that the company and its artists “recognize the significant potential and progress of artificial intelligence,” but using copyrighted content for AI training, development, and commercial applications without authorization is “denying Sony the right to appropriate compensation.”
Sony Music Group harshly warns these technology companies that, given “the nature of their operations” and “information already published in AI systems,” the company has reason to believe that these companies (or their affiliates) may have used Sony content without authorization. Sony asked these developers to provide details of the content they are using by next week.
The smoke has already risen
In the context of Sony Music Group, these AI companies infringe not only music melodies, but also music album cover art, lyrics, and metadata.
Copyright issues are already a major cause of commercial disputes at AIGC. Hollywood screenwriters and actors went on strike last year to demand that their work be protected from artificial intelligence. As the latest batch of “Wensheng Audio” startups begin using technology to generate finished music albums, the tension between artists and the AI industry continues to escalate.
Similar to Sony Music, Universal Music Group has also taken an active stance to defend its rights. In addition to removing all copyrighted music from third-party short video platforms by using the “nuclear option” to force them to promise that “the development of AI will protect the financial interests of artists,” it is also currently going to court with the well-known AI startup Anthropic.
In April of this year, Warner Music Group CEO Robert Kyncl also stated during a congressional hearing that he supports the introduction of federal legislation by the United States to regulate AI companies' disregard for copyright.
“We must ensure a strong free market licensing system, use copyrighted materials to train artificial intelligence models, and provide strong legal protection for name, image, and voice rights,” Kyncl said in testimony.