Is the Future of Music in the Hands of Robots?

Music is innately human—it has been a constant of the human experience for millenia and elicits a broad spectrum of irreproducible emotions. It is almost impossible to imagine music being produced by anything other than people. However, artificial intelligence is making waves in the music industry, challenging human creativity. While many consider AI only for its practical uses, it has expanded into the realm of music and art optimization. It may initially seem intrusive, or even dystopian—that mindless simulations are encroaching upon our domain—but an argument can be made that the presence of AI might not serve as an obstacle for human expression, but rather spark a new era of creativity and push music into unexpected directions.

The technology behind music AI utilizes neural networks, or software that attempts to mimic the human brain. A music producer sends pre-composed music to the neural network, which then recognizes melodic, harmonic, rhythmic and timbral patterns to generate new music. The neural network has grown to hold such a multitude of songs over the past decade that it is challenging for humans to monitor their extent of data.

One company that has become a masthead for the public consumption of AI-generated music is OpenAI. Through their own public neural network named Jukebox, they are able to spontaneously generate new music by modeling uncompressed audio, a departure from developing music symbolically, by recognizing elements like piano chord progressions. Jukebox’s pivot to raw waveform synthesis grants computers the ability to replicate the human voice while developing unique lyrics—a novel accomplishment inaccessible by older AI models.

However, with this adaptation, Jukebox can easily become a copyright disaster. Say, for example, AI is trained to mimic the style of Taylor Swift: would she own any rights to the generated music? The copyright issue might restrict the potential spread of AI, as the music industry has already seen artists remove a synthesized Jay-Z song from YouTube. Although copyrighting an algorithm is challenging, this issue will only grow worse with the development of AI, and it is only fair that artists who own the creative property to the original piece receive some sort of compensation. Even though the legality of AI-generated music is questionable, it remains an innovative development that pushes the boundaries of the music industry.

Alternatively, some artists have already begun to utilize AI in their own creative processes. In 2016, musician Alex Da Kid (Alexander Grant) used AI to analyze contemporary hit songs, social media data, and top-grossing films to compose “Not Easy,” using popular pitches and note sequences. He reached number four on the iTunes Hot Tracks chart within 48 hours of the song’s release.

After seeing the success that AI artists have brought to the music industry, AI has begun to be used for finding the next big hit. AI has been helping large corporations like Warner Music Group and Apple find new artists through its abilities to quickly shuffle through music. Applications like Musiio feature an algorithm called Hit Potential Algorithm, which is not only able to categorize and assess new music but also measure how likely it is to reach commercial success. Musiio can determine tracks with no plays, no release date, and no social media marketing as potential hits. Although this process is less effective than that of a professional talent scout, its speed is unmatched, meaning more artists can be discovered. Apps like Musiio are best used in tandem with the human ear to pre-emptively sort through 40,000 tracks a day, saving time for those searching for potential stars.

As this technology continues to develop, it will create new accessibility to the music industry. Up-and-coming artists may be able to create professional-sounding music with merely a phone instead of devoting money to expensive royalties or a recording studio. Music AI could also benefit established artists like French musician SKYGGE, who created the fully-AI assisted album “Black is the Colour” (2019), which saw their style shift from French pop to traditional American folk.

Though there have been positive effects on musicians from using Music AI, there still are concerns for its future. One is whether or not the optimization of music with AI will shift the focus of music away from art and onto corporate bottom lines, which can carelessly push human creatives out of business in favor of efficiency. Currently, AI tunes are clearly discernible from human-created music, but even if new technology was able to create music independent of its base of human art to draw from, it would never sufficiently replace human artistry. AI holds no guiding values, fresh perspectives, or cultural analysis, meaning that unlike human music, it is devoid of some of music’s greatest communicative utilities, such as artistic intent, political commentary, or emotional connection. For now, that is.

Music AI is quickly changing the environment of the music industry, paving new paths to the music industry, and giving composers complex tools to enhance their creativity, and finding diamonds in the rough of music created each day. Advances in AI show no signs of slowing down, so the future of music remains exciting with the exploration of new sounds and techniques that have never existed before.