The Grammys set new rules, and the music industry is embracing AI?

Source: Sound of Music, Author: Wan He, Editor: Fan Zhihui

Image source: Generated by Unbounded AI tool

Musicians who hate AI creation can breathe a sigh of relief.

Previously, an AI-generated song "Heart on My Sleeve" based on the voices of Drake and The Weekend became popular overseas and is considered to be the first out-of-the-box AI music. Although the song was later removed from major platforms due to pressure from Universal Music, it seems that the controversial AI replacing human music creation is becoming a reality. There is even concern that the next step for AI music might be to win a Grammy.

Recently, the Grammy organizer National Academy of Recording Arts and Sciences (National Academy of Recording Arts and Sciences) updated a series of selection rules, giving the answer in terms involving AI music: entries can use AI elements, but Entries generated entirely by AI are not eligible for any awards. In other words, at least from the perspective of the new Grammy regulations, AI cannot replace humans.

As soon as this move came out, although it was opposed and questioned by many netizens, it has to be said that Grammy's attitude is indeed in line with the current attitude of the European and American music industry towards AI creation. However, facing the wave of AIGC, the music industry is actually acting silently, choosing to join if they can't beat it.

**Can AI win the Grammys? **

In the face of the accelerated development of AI technology, it is not impossible for human creators to worry, and the industry represented by Grammy can be regarded as a reassurance.

The rules state, "The Grammy Awards honor creative excellence. Only human creators are eligible to be submitted for consideration, nominated for, or won for a Grammy."

**Although entries may contain AI assistance, the Grammys stipulate that the human creator part of the work must be "meaningful" and relevant to the competition award. **For example, if it is a songwriting category, the human creator needs to have a clear contribution to the music or lyrics, and if it is a performance category, then the human musician also needs to have a "meaningful" participation.

Also, to be nominated in the Album of the Year category, creators need to contribute at least 20% to the album to be eligible. It is worth mentioning that there was no such regulation before. In the past, any producer, songwriter, engineer, or artist featured on an album could be nominated for Album of the Year, even if that person's contribution was minimal. Now that AI is in power, although record companies are particularly worried about AI infringement, Harvey Mason Jr., CEO and president of the Grammy organizer, emphasized in an interview that the industry needs to accept AI and establish standards to adapt to the impact of artificial intelligence on the art world. and impact on society as a whole.

For this new regulation, most netizens expressed doubts or opposition. Bindu Reddy, co-founder and CEO of Abacus.AI, believes that this is a kind of discrimination against AI and an escape from competition by human beings. It works."**

Some netizens on the American social news website Reddit believe that this rule is not easy to implement. With the frequent emergence of various AI-assisted creation software, AI can assist in everything from lyric writing, composition, singing and even mixing and arrangement. It has become almost impossible to completely exclude AI. Because the boundary between human music and AI music is quite blurred and difficult to identify, "What if I use ChatGPT to write lyrics? How do you know if I wrote it? The point is, how to draw this line and how to execute it?"

** Some netizens worry that AI technology will become another monopoly advantage of mainstream artists. **

After all, mainstream record companies have the funds to develop and invest in more cutting-edge AI technology to improve the presentation of music, thereby widening the gap with independent musicians. This has actually happened. Although the quality of music has nothing to do with technology, it has to be admitted that mainstream record companies and artists are quite active in exploring AI technology, and have indeed created some of their own advantages.

For example, just this month, Paul McCartney, the former backbone of the Beatles, announced that he would use AI technology to extract the voice of John Lennon to make "the last Beatles record." Previously, HYBE, which acquired the AI voice company Supertone, also launched the artificial intelligence artist MINDNATT this month, which is equivalent to the digital clone of the famous Korean folk singer Lee Hyun, and helped Lee Hyun's song "Masquerade" to be released in six foreign languages. There are even female vocal versions of the song. "Masquerade" surpassed 750,000 views on YouTube in just 7 days, and was also thanked by many Spanish-speaking fans.

After listening to the Chinese version of the music, I feel that although the Chinese pronunciation is indeed relatively standard, it sounds particularly emotionless, and some accents still feel unnatural, like chanting rather than songs. If we want to achieve a more natural effect, I believe that the staff will need to work hard on different languages and cultures in terms of lyrics and tone expressions, but such rapid progress is disturbing enough.

As Harvey Mason Jr. said in an interview, "It is not advisable to be disrupted by AI music and not face it. I am a little hesitant about the unknown impact of AI in the next few months and years. And worry. But I know full well that this is bound to be a part of the music industry, the art world, and society as a whole.”

Music industry has embraced AI music with concern

The attitude of the Grammys is also consistent with the attitude of the music industry: ** Accept AI as an auxiliary tool for creators, not a substitute for human music. **

This year, Universal Music has repeatedly emphasized the copyright issues that will be caused by the development of generative AI. After someone used the voice of its artist Drake to make a song and became popular, it immediately asked the platform to remove the song. The Recording Industry Association of America (RIAA), which represents the interests of record companies, has also united dozens of music industry organizations to establish a new industry alliance "Human Artistry Campaign" (Human Artistry Campaign) to supervise the use and development of AI without infringing on human culture and art. .

Nevertheless, Universal Music is not resisting this technology, but is actively using generative AI as a tool to assist the company's business development.

In the past two years, Ingrooves Music Group, a label of Universal Music, has been committed to automating AI-assisted marketing activities, identifying high-value audiences, and predicting the behavior of high-value social media audiences, and has obtained three patents in this regard. Last month, Universal Music also established a strategic partnership with AI sound startup Endel to create "AI-assisted, musician-driven functional music." Previously, the company has received investments from music-related companies such as Warner Music, Sony, AVEX, and Royalty Exchange.

At the same time, the music companies represented by the three major records have also invested in or acquired many AI creation platforms. Last year, Universal Music invested in Soundful, an AI-assisted music creation platform; at the same time, Warner Music invested in Lifescore, an AI-assisted music creation platform, and Sony Music launched Flow Machines, an AI-assisted creation platform.

Recently, TuneCore, a platform under Believe, also announced to block music generated entirely by AI. According to Believe, Believe is working with AI companies to monitor AI-generated audio tracks with 99.9% accuracy; at the same time, Believe is also exploring the realization of AI-generated music and feeding back its used copyright content (such as voice performance ).

Although the whole world wants to take the lead in the development of AI, since AI model training requires a lot of human creation content, its wanton development can easily infringe the interests of copyright owners, so risk management is equally important.

**On the legislative level, the attitudes of various countries are not the same, or even contradictory. **

Among them, the EU is perhaps the most conservative. As the earliest region to regulate AI with legislation, the European Parliament voted to pass the authorization draft of the "Artificial Intelligence Act" (AI Act) this month, which will come into force in 2024, becoming the world's first artificial intelligence legislation. The goal of the decree is to "ensure that the AI systems used in the EU are safe, transparent, traceable, non-discriminatory, and environmentally friendly." It will regulate various AI technologies to varying degrees, but it is unclear to what extent it will be implemented.

** In sharp contrast to Europe's cautious approach is Japan. **Here, from the central government, down to localities and enterprises, all have shown a strong interest in ChatGPT. In April of this year, Japan’s Minister of Education, Culture, Sports, Science and Technology Nagaoka Keiko made it clear that the Japanese government believes that using copyrighted content to train AI does not violate copyright law, even for commercial purposes. This year, Yokosuka City in Kanagawa Prefecture has tried to fully introduce ChatGPT to local governments for official business.

In contrast, the UK and US are trying to find a balance between risk and opportunity.

The UK released a white paper titled "A pro-innovation approach to AI regulation" (A pro-innovation approach to AI regulation) in March this year, but it only proposed some guiding principles for regulators. Although the United States released the "Artificial Intelligence Risk Management Framework 1.0" to provide a management framework for relevant institutions to design and manage trustworthy artificial intelligence, it has not yet proposed comprehensive legislation related to regulation; until this month, the US Congress seems to have finally Aware of its risks, the two parties jointly proposed the establishment of a national AI committee to carry out relevant legislation.

Although the legal attitudes of various countries are still ambiguous, financing in the field of AI music is already in full swing.

In the last month alone, there have been more than 5 AI music-related investment and financing events, including: music technology startup Cyanite acquired AI sampling platform Aptone; generative AI-driven music startup BIDBOX.IO has completed a round Seed financing, received $7 million in financing from Forerunner, super{set} start-up studio and Ulu Ventures; AI music streaming platform WAVs AI received $20 million in financing from Canadian venture capital Regal Investments; AI splitting company AUDIOSHAKE from PeerMusic A number of music companies have raised $2.7 million in seed round funding.

As mentioned above, many mainstream musicians are also actively embracing AI technology. The Canadian singer Grimes, who is always at the forefront of technology, not only did not worry about his voice being infringed, but also took the initiative to cooperate with the AI platform Elf.Tech to allow all platform users to use Grimes' voice on the condition that they share 50% of the recording royalties of the song.

Elf.Tech main interface

Today, Grimes has reached a cooperation with TuneCore, an independent music distribution platform, so that after musicians use Elf.Tech to create, their works can be distributed to all major streaming media platforms with one click.

Conclusion

If the government and music companies still need to find a balance between risks and opportunities, then musicians have reasons to try AI creation tools, whether from the perspective of curiosity or fear.

In fact, the percentage of musicians already using AI may be much higher than we think. According to a study released by Ditto Music in April this year, Ditto Music found among more than 1,200 surveyed users that nearly 60% of musicians are already using AI, including making album covers, mixing, and creating.

But while AI is able to share various tasks, it also replaces the work of some people. "Mastering engineers will be the first to leave, then mixers," said YouTube music personality and senior producer Rick Beato, who expressed concern about the recording industry. He also predicted that artificial intelligence mixing/mastering tools will Can imitate anyone's style.

With these AI tools that improve all aspects of music production, it has become difficult to identify or eliminate whether AI is involved in the creation. In addition, the use of AI-assisted creation has gradually become mainstream. Without the government's mandatory intervention, it is almost impossible to deny AI creation. possible things.

**From this point of view, the Grammy’s new regulations may not be ambiguous, but a firm statement that no matter how AI technology develops, the art of music must still be the privilege of human beings! **