Text | Music Precursor, Author | Li Qinyu, ed Fan Zhihui
On May 16, Sony Music released a statement on its official website, announcing that Sony Music will withdraw from AI training, prohibiting any company from training AI models with its content without authorization, including but not limited to melodies, lyrics, audio recordings, audio-visual recordings, illustrations, portraits, etc.
According to the BBC, Sony Music Group has so far issued warning letters to about 700 AI companies and streaming platforms. In the letter, Sony Music recognises the "great potential and progress" of artificial intelligence. However, the unauthorized use of Sony Music's content in the training, development, or commercialization of AI systems deprives them of control and revenue over the use of their content.
Among other things, Sony Music Group asks each recipient to provide details of the songs it uses to train the AI system, as well as how to access the songs, ask for the number of copies of the songs, etc.
Although Sony Music Group did not specify which 700 companies sent the letter, and the full version of the letter has not been made public, according to people familiar with the matter, it includes OpenAI, Microsoft, Google and many other large companies.
According to foreign media reports, in addition to AI developers, SMG also sent separate letters to Apple and Spotify, demanding that streaming platforms take the best measures to protect artists and songwriters, circumvent illegal training models by unauthorized AI companies, and ask them to update their terms of service to prohibit mining and scraping their content.
To put it simply, the letter is a warning to some 700 AI companies and streaming platforms, either admitting that unauthorized infringement has occurred or, if not, providing details of AI training.
Sony Music takes the lead in "quitting the group chat"
Sony Music Group is doing this because of its market position as one of the three major records, as well as the increasingly standardized laws and regulations.
In March, the European Union passed the Artificial Intelligence Act, which requires providers of general AI models to disclose "sufficiently detailed" summaries of what they use to train them. According to a principle introduced by the AI Act from the EU's 2019 Copyright Directive, AI developers must obtain permission to use copyrighted material if the copyright owner opts out.
Despite the industry's opposition to the provision at the time, the U.S. Music Rights Agents Association (NMPA) stated in a 2023 submission to the U.S. Copyright Office that "the opt-out system fundamentally undermines copyright protection by shifting the burden of obtaining a license to users." ”
It is understood that Sony Music Group is also the only company among the three major records that did not participate in the new features of YouTube's "Dream Track", but the company has not completely shied away from artificial intelligence.
Last summer, Sony Music hired its first executive vice president of artificial intelligence, Geoff Taylor, to launch the AI project "Metallic Spheres In Colour" by the electronic band The Orb and veteran band Pink Floyd in November, inviting fans to re-engage in creating the music and cover for the new album using AI tools.
At the U.S. Senate Artificial Intelligence Forum last November, Dennis Kooker, president of sales for Sony Music in the U.S., also made clear the big music company's stance on artificial intelligence and copyright law in a speech.
Dennis Kooker has publicly discussed Sony Music's efforts to remove unauthorized AI content from online platforms, including nearly 10,000 individual takedowns that the company has announced. In addition, Dennis Kooker made it clear that opposition to Google's AI developers "using copyrighted material to train AI models should fall within the fair use exemption of copyright law." ”
According to Dennis Kooker, even if the material is copyrighted, if it is allowed to be used by AI companies, it means that "some companies can usurp all the value generated by the creative department without permission and build a huge business on top of it, without paying any money to the creators involved." ”
It can be seen that Sony Music's strategy in the field of artificial intelligence is more inclined to develop tools that can enhance creativity and protect content, rather than participating in experiments that may involve greater copyright and moral risks. The letter to about 700 AI companies and streaming platforms also reflects Sony Music's determination to create a win-win situation for copyright owners, artists and AI technology providers.
The AI music copyright dispute has never stopped
In recent years, the application of AI technology in the music industry has long been common, both at home and abroad.
According to a BPB (Music Production Forum) survey of 1,500 music producers last year, about 36.8% of producers already use AI tools in the music production process, while 30.1% plan to try using AI tools. According to MBW, in 2023 alone, 170 million songs will be created using AI and uploaded to music streaming platforms such as Apple Music, Spotify, and Amazon Music.
Whether it is AI music tools such as Suno, AIVA, and Boomy abroad, or AI music tools such as TME Studio and Xstudio in China, they provide diversified functions in songwriting, mixing, and music generation.
As a result, copyright disputes and lawsuits are on the rise. According to information from Sustainable Tech Partner, an overseas digital media platform, from 2023 to the present, there have been 40 AI music-related disputes, of which 8 involve lawsuits.
In April, the Artists Copyright Coalition and about 200 artists, including Billie Eilish, Nicki Minaj and Katy Perry, published an open letter calling on digital music developers to "stop using artificial intelligence to infringe on and disparage the rights of human artists."
Not long ago, TikTok and Universal Music Group jointly announced that they have reached a new copyright licensing agreement. One of the most important agreements is the commitment of the two parties to "work together to ensure that the development of AI across the music industry will protect human art, and that the resulting economic benefits will go to those artists and songwriters." ”
Also in an embarrassing situation is OpenAI.
In an interview with The Wall Street Journal, Mira Murati, the company's chief technology officer, was vague about the source of Sora's training data after launching its popular text-to-video generation model Sora in March last year.
In May this year, according to CNBC, Google's CEO Sundar Pichai said that if it is determined that Microsoft-backed OpenAI relies on YouTube content to train artificial intelligence models capable of generating videos, Google will take legal action to solve the problem.
Since the birth of AIGC technology, the copyright dispute over its training content has never subsided, and now it has really crossed the cake of creative industry giants. Sony Music's official warning this time also added another fire to the AIGC's copyright defense war.
The tools provided by AI for artistic creation have long been more than a threat to creators and copyright owners, which is not like the evolution of painters putting down their brushes and picking up hand-drawn tablets, but more like subverting the future of the industry.
From legislation to technology, how can the value of human art be maintained?
Since December 8 last year, the European Parliament, EU member states and the European Commission reached an agreement on the Artificial Intelligence Act. On March 13 this year, the European Parliament passed the Artificial Intelligence Act with 523 votes in favor, 46 against and 49 abstentions.
The Act is the world's first comprehensive legislation for the widespread use of AI, aiming to ensure that AI systems deployed and used on the EU market are safe and consistent with fundamental rights and EU values. The bill includes strict regulations on high-risk AI systems, new governance requirements for general AI models that could have a significant impact on society, and the creation of a dedicated AI office to oversee and enforce these regulations.
On the other hand, the U.S. state of Tennessee has also recently passed the Secure Portrait Sound and Image (ELVIS) Act to protect the entertainment industry from the misuse of artificial intelligence. Unlike the Artificial Intelligence Act, this new bill focuses on protecting an individual's voice and image from the unauthorized use of AI, and can be held accountable through civil litigation for the unauthorized use of an individual's voice or image.
On a technical level, in May this year, French audio technology company Ircam Amplify launched a new tool, AI-Generate Detector, designed to identify music created using artificial intelligence, which can scan up to 5,000 music tracks in a minute with an accuracy rate of up to 98.5%.
Nathalie Birocheau, CEO of Ircam Amplify, said: "Our AI-generated detectors are a game-changer. It will provide transparency where it is not transparent, giving users a strategic advantage and helping them make informed decisions about their music content. ”
AI tools, anti-AI tools, and the promulgation and enforcement of relevant laws are starting a protracted battle.
On the one hand, we should actively embrace changes in the industry, try more and apply AI tools to prevent elimination, and on the other hand, we should always pay attention to changes in the wind direction and protect the current rights and interests from being infringed.
epilogue
Science fiction movies are coming to life.
When human beings use tools to make a clear line between them and all things, they gradually bring a sense of God into the rapid development of creation.
Whether it is intelligent robots or artificial intelligence technology, it marks its "artificial" characteristics with the brand of "human". The birth of this silicon-based life has given mankind the feeling that God created man, and its creators have also shown the dual longing and worry of human beings for the development of science and technology - on the one hand, it is a beautiful vision of human-centered and high-tech will always serve every human being equally like "S1m0ne" and "Artificial Intelligence", and on the other hand, as "The Matrix" and "Westworld" are worried about, they are worried that their highly developed wisdom will eventually subvert human society and cause inevitable threats.
However, the future that has been discussed in science fiction for decades has crept in, and artificial intelligence has rudely intruded into our daily lives. The wind of science and technology hitting traditional industries has finally blown into the field of human art.
This is contrary to the previous optimists' expectation that AI will handle the tedious and repetitive daily life and leave creativity to humans, and what is even more worrying is that it will still take a long time for the institutional level such as laws and regulations to be reasonably regulated.
How to minimize the negative impact of artificial intelligence technology and prevent the music industry from becoming a chaotic and chaotic primeval jungle in the future will become an inevitable industry topic for a period of time in the future.
After all, in the Bible, even Adam and Eve, who God created man and put it in the Garden of Eden, could not escape the temptation of the serpent, a symbol of wisdom, to eat the forbidden fruit and rebel.