Music, as many other creative fields that are also highly innovative and competitive, has seen the rise of artificial intelligence (AI) and machine learning and explored ways to take advantage of them.
Long existing music platforms like Spotify and Apple Music have already demonstrated new ways to listen to music and engage with established or emerging artists. As a data-driven on-demand music platform, Spotify takes advantage of its users’ data to recommend music and create personalised playlists. Meanwhile, after Apple broke up iTunes and introduced a number of standalone media apps, it continues to add more music content to Apple Music, including livestreams, new DJs, pre and post-album coverage.
However, on-demand platforms are only the tip of the iceberg of the music industry’s strive for innovation. As AI and machine learning develop, these new technologies are enabling artists and music producers to produce music in wholly different ways to the established industry workflows.
London-born musician and artist Reeps One (Harry Yeff) experiments with innovative vocal practices alongside the creative representation of physics and sound. In 2018, he presented the first ever beatboxing battle between him and an artificial intelligence (AI) opponent. To do this, he partnered with programmer and artist CJ Carr to develop a deep-learning program, taking a lecture Reeps had given in Sweden and transforming it into a completely new string of sounds.
His specialisation in oral percussion and performance has generated a global online following, gathering over 50,000,000 views, and an established reputation as a key pioneer in the field of experimental vocalism. Internationally recognised as a leading artist of the New School Beatbox Scene, Reeps One will be discussing his work in AI and music as a speaker at Beyond.
Similarly, London-based AI startup and ‘home for virtual entertainment’, Auxuman has created a group of AI artists, each with their own personalities, styles and music genres: Yona, Mony, Gemini, Hexe and Zoya are to release full-length albums every month. The company is dedicated to “building the next generation of virtual entertainers”. According to British-Iranian music producer Ash Koosh, “the music is created through engines that create the words, melodies and a digital singing voice”. The machines that make the music with lyrics are “trained on articles, poems and conversations related to the subject or a song.
However, as AI continues to develop, it is also increasingly able to mimic artists’ style, which poses questions on originality and copyright policy. For example, if an AI is exclusively trained on Beyoncé and is then able to imitate the sound, is the artist owed anything? Another example are the data scientists at CableTV.com who fed all of Taylor Swift’s lyrics into a neural network. The resulting song, The Last Word (Whoa, Whoa-Ah-Oh), includes both realistic Swift details and AI elements.
A recent article on The Verge suggests that currently if a copyright claim was filed against an AI, it would be difficult to prove an algorithm was trained on the song or artist, on whom it allegedly infringes.
Therefore, while AI might prove a valuable tool for artists in music production, it also complicates the industry’s established legal and ethical policies, and challenges current levels of creativity.
It’s clear that experimentation with AI and data in music is further challenging the industry, but it is also opening up exciting new channels that are pushing the boundaries of creativity. Discussing the implications and impact in more detail, Audio & Visual Solo Artist Reeps One, CEO of Phrasee, Parry Malm and Scriptbook Founder & CEO, Nadira Azermai, will be at BEYOND to explore how AI is helping artists to become more creative and extending human abilities, in the ‘Just a Member of the Band’ conference session.
About BEYOND
The challenges and points raised in this article were one of the key themes to be addressed at the BEYOND conference, in Edinburgh on 20-21 November 2019. Exploring the impact of AI, Machine Learning and Data on the creative industries, the two-day programme offered lively debate, keynote sessions, short-talks, panel discussions and exhibitions showcasing the UK’s excellence in research-driven creative innovation, plus networking.
© Main photo by Franck V. on Unsplash