At 1pm on the 15th, a new artist announced his existence to the world. The digital single ‘Masquerade’ begins with a synth sound reminiscent of sirens, and a strange tension continues throughout. In the accompanying music video, <듄> A space reminiscent of the mysterious universe within unfolds.
The one who pulled off the veil that day was MIDNATT, an artist belonging to Big Hit Music, a label under Hive. He is the alter ego of singer-songwriter Lee Hyun of the group 8eight, and is the protagonist of the new concept project ‘Project L’, which was announced by Chairman Bang Si-hyeok of Hive through an interview with US Billboard magazine the last month.
AI appears in the human song
The reason why Midnat is special is that it was born as ‘a combination of technology and music’. In “Masquerade,” a woman appears in addition to the male lead singer, giving it the feel of a duet song. However, this woman does not exist. Hive applied the ‘voice design technology’ of Supertone, an AI audio company it acquired in January, to create a female voice using Midnat’s voice.
The technology was also used to launch ‘Masquerade’ in six languages (Korean, English, Japanese, Chinese, Spanish and Vietnamese) for the first time in the world. When Midnat sang in six languages, it was done naturally through Supertone pronunciation correction technology.
Chroma and LED augmented reality (XR) systems were simultaneously applied to the music video which created a world that didn’t seem to exist. In the music video, the entire section except for the forest location is the background based on XR.
Jung Woo-yong, CEO of Hive IM, which designed the project, said at a press conference held at CGV Yongsan in Seoul that day, “We applied the technology very carefully so as not to harm the identity of the artist.” and we are proud to expand with our voice design technology.”
aespa 에스파 ‘Welcome To MY World (Feat. nævis)’ MV
The expansion of the music world using AI technology continues. In addition to Midnat, many artists are trying to combine with AI in many ways, such as content.
3rd mini album <마이 월드>A representative example is the group Aespa, who released the On the 2nd, before the release of the album, ‘Welcome to My World’, which was released in advance, there was ‘Naives’, a supporting character in the aespa worldview. Nivis was responsible for the humming at the beginning of the song and the chorus in the middle. They were praised for harmonizing with the members’ voices and creating the atmosphere of the song. It is known that Nivis’ voice was created from scratch by combining the voices of several voice actors.
At a comeback press conference on the 8th, member Karina said, “I’ve danced with AI and filmed a music video, but this is my first album.” He said, expressing his impressions of breathing together. Espa introduces the concept of a total of 8 members, including 4 members in the real world and 4 avatars in their virtual world.
The agency, SM Entertainment, is preparing for Navis’ debut as a singer. For this purpose, development work for virtual humans is known to be underway. ‘Welcome to My World’ was initially planned to be a solo song by Navis, but at the request of the Espa members who liked the song, Navis only took part as a feature song.
The legal and ethical dilemma of AI music
The use of AI in popular culture is expected to gradually expand its scope. But is there only joy in the world of music created by AI?
Recently, a number of videos have become a hot topic in social networking services (SNS). Videos of pop star Bruno Mars singing “Hype Boy” by K-pop group New Jeans or Canadian singer-songwriter The Weeknd singing “Cupid” by another K-pop group Fifty Fifty garnered hundreds of thousands of scenes. The K-pop covers of world famous stars attracted the curiosity of many.
Bruno Mars and The Weeknd never sang these songs. Both videos are called ‘AI overlays’. After implementing the voices of two people using AI voice conversion software, it was made as if they were singing a specific song. The high quality and similarity of voices gave users a “goosebumps” response.
Hype Boy – Bruno Mars (Original by Newjeans) (AI COVER)
Similar cases follow one after the other, domestically and internationally. The voices of countless artists such as Rihanna, Ariana Grande, and Drake were created with AI. In the UK last month, even the band ‘AIsis’ appeared, who released an album by doing the voice of Liam Gallagher, the vocalist of the legendary band Oasis, with AI.
Trouble arises at this point. Most of the AI cover videos were produced without the permission of the artist, and some of them lead to monetization. The industry is concerned that there is a possibility of copyright infringement and defamation if AI coverage were to be done indiscriminately. An official from the Korea Music Copyright Association said, “Organizational and legal arrangements related to creative works such as music created or participated in by AI have not yet been properly maintained.” There could be problems,” he said.
As much as a particular person’s voice can be implemented in the same way, there is a possibility of abuse. As a fake technology, there are concerns that it could be used for digital sex crimes.
Jeon Chang-bae, chairman of the International Association for Artificial Intelligence and Ethics (IAAE), said, “It is welcome to create content using AI in the fields of entertainment and media,” but he also added, “Since the generative AI tools can be used by anyone, there are ethical issues when implementing them without the consent of real people,” he noted.