As OpenAI, Google, Adobe and other major tech corporations rush to roll out technology that uses generative AI to make music, competition in the AI music market is expected to intensify. Observers say the era has arrived in which, if you write a one-line description of the feel you want, AI can generate a highly polished song in as little as 10 seconds, fundamentally changing how music is produced.
According to the AI industry on the 4th, OpenAI is developing a new tool that generates music when given a one-line command in text or speech. IT outlet The Information reported that the tool, which is based on text and audio prompts, is expected to be used to add background music suited to videos or to layer guitar accompaniment onto a vocal track. To secure training data, OpenAI is working with students at the Juilliard School in the United States to annotate sheet music.
However, it is said to be undecided whether it will be able to generate a complete song from start to finish from a text description of the desired music, as AI music Start - Up Suno does.
If OpenAI also releases a music-generating AI tool, it is expected to acquire full-stack content-generation AI capabilities spanning text, images, video and music. AI music created by the new tool is believed likely to be integrated with Sora, the AI video-sharing social media platform OpenAI recently launched.
With OpenAI joining the fray, the battle for dominance in the AI music market, which includes Google and Suno, is also expected to heat up.
The AI music market is in fact growing rapidly. According to market research firm Grand View Research, the generative AI music market is expected to grow about sixfold from $440 million in 2023 (about 620 billion won) to $2.7947 billion in 2030 (about 4 trillion won). Grand View Research said, "As AI allows anyone to make music, the barrier to entry for music production has fallen, and by automating repetitive tasks such as mastering and editing, AI has made it possible to reduce the time and expense required for music creation."
Suno is an AI music Start - Up with annual sales of $150 million (about 210 billion won). If you enter a prompt such as "a calm song that heals you on your way home from work," its music-generation AI service completes melody and lyrics in 10 seconds. According to Bloomberg News, Suno's corporate value is estimated at about $2 billion (about 2.86 trillion won).
Among Big Tech corporations, Google unveiled its music-generation AI model Lyria in 2023 and released a second, improved model in April this year. Lyria is used to supplement or transform music based on an artist's melodic ideas. Google said, "Lyria is designed on the premise of collaboration with musicians and other music creators, and it can generate audio using text prompts."
At Adobe Max, its annual creator conference held last month, Adobe unveiled Soundtrack Generation, which creates custom background music for video from a single line of instructions. If you enter a prompt such as "an electronic ambient-style track with a peaceful and contemplative mood for a travel documentary," an instrument-based soundtrack suitable for use in films, drama, and ads is created within 10 seconds. Even without separate instructions from the user, AI can analyze the video and recommend a matching musical mood, tempo and energy level.
Negative perceptions and copyright issues surrounding AI music are challenges the AI industry must overcome. Major record labels argue that AI models used copyrighted audio and lyrics without permission during training. Last year, the world's three major record labels—Warner Music Group, Universal Music Group and Sony Music Group—filed large-scale copyright infringement suits against Suno and another AI music Start - Up, Udio.
As so-called "AI artists" (virtual musicians created by AI) rack up millions of plays on streaming platforms such as Spotify, concerns have emerged that AI music will gobble up streaming revenue and take jobs from singers.
This year, Velvet Sundown drew attention as a breakout rookie after logging more than 1 million plays on Spotify. But controversy erupted when it was revealed that everything from their music to promotional images and band concept had been created with Suno AI.
On the 18th of last month, U.S. singer Xania Monet became the first "AI singer" to make the Billboard radio chart. The singer also signed a record deal worth $3 million with a label. CNN reported, "With more than 146,000 Instagram followers alone, Xania Monet is a representative example showing that consumers are becoming increasingly comfortable with AI artists." Monet is an AI singer created by Mississippi-born poet and lyricist Telisha Jones and is produced in a way that has Monet "sing" with help from Suno AI based on lyrics written by Jones.
In a report published in September, music streaming service Deezer said, "Twenty-eight percent of the tracks posted on Deezer were AI-generated music," adding, "As recently as April, that figure was only 18%, and it is rising rapidly."