Generative AI – tips, trends & consequences

Generative AI has entered the scene in a big way – and its development continues at a pace that few could have predicted. But what does this mean for you as a songwriter and composer? How can you use AI when creating music? How do music publishers and Koda handle your music rights in this brave new world? And last but not least: gazing into the crystal ball, what are the future outlooks for technological development and music?

This article is a snapshot of the state of things as of June 2024. At the time of writing, developments in the field of generative AI are moving incredibly fast. As a result, changes are to be expected regarding many of the topics addressed in the text.

In mid-June, a seminar held by Autor, DKF, DPA, Musikforlæggerne and Koda featured many invited guests: DJ Noize and Vita Dosa demonstrated how they use AI in their songwriting and production. Christian Fløe Svenningsen from GL Music enlightened us on how AI influences their work, while Sofie Hvitved, who is a senior researcher at the Copenhagen Institute for Futures Studies took a look into the future as seen through the lens of AI. Koda’s Tech Scout, Nicolai Franck, talked about how Koda endeavours to protect the interests of songwriters and composers in the age of generative AI, and about how Koda manages songs where generative AI has been used in their creation. Read on and see what AI means for you.

How Koda works with AI

Koda’s Tech Scout, Nicolai Franck, talked about the steps Koda is taking to protect the rights of songwriters and composers in the age of AI. Koda currently focuses on four main areas:

Copyright and policy: AI companies must ask permission and pay for the use of copyrighted songs and lyrics to train their AI models.
Work registration: Koda has developed clear guidelines on how to register songs where AI has been used in the songwriting process. We welcome new technology, but at the same time we protect what has been made by human hands and minds.
Legislative regulation of AI-generated music: Legislation must establish a framework for the use of fully AI-generated music. Remuneration for such use must be made to the original rightsholders whose works form the crucially important basis enabling the AI companies to build and offer their AI music services.
Licensing: AI services wishing to use works created by Koda members must enter into licensing agreements to do so.

‘At Koda we welcome new technology, but at the same time we want to ensure the best possible conditions for human-made music and culture. We are working hard to ensure that Koda’s members receive fair payment when tech companies use the members’ copyrighted songs to train the AI models that enable them to offer AI services,’ said Nicolai Franck and continued:

‘Ultimately, the task of defining clear rules that will safeguard a sustainable music scene in the age of AI will be a political one. It is a task that must be solved jointly, which is why Koda collaborates with other organisations in Denmark and internationally to attract the politicians’ attention to the issue.’

The use of generative AI in music can be divided into four overall scenarios:

Voice cloning and deep fakes: AI can imitate voices, including those of well-known artists. When done without permission, such imitation is problematic and controversial. The technology can also be used to recreate the voices of dead artists or as a creative tool where, for example, you can sing in other languages than your own using a clone of your own voice.
AI-assisted songs: AI can act as a co-writer that generates ideas and initiates creative processes. It can also help fine-tune the artist’s own ideas.
100% AI-generated songs: Music created entirely by AI is becoming more common and harder to distinguish from human-made music. This could potentially become a major competitor to human-made music, and such competition will be unfair.
Stem splitting and sound transformation: AI can separate out individual tracks in a recorded song, as was demonstrated with the new Beatles song ‘Now and Then’ in late 2023. The technology can also be used to transform the sound of individual parts of a song.

See more about how Koda works with AI.

AI from a music publisher’s perspective

Christian Fløe Svenningsen from GL Music highlighted three main areas:

AI artists: AI creates new ways of presenting artists, such as Naevis and Aespa. Christian noted that ‘we stand on the brink of a new adventure, not at the edge of an abyss’.
AI tools: AI-generated voices are used for pitches that are later performed by human singers. This makes it possible to present ideas quickly and efficiently.
A&R bias and AI: Tools like Disco use autotagging to classify music, making it easier to find and sort music. AI is a valuable tool for increasing efficiency and achieving faster results. Christian described AI as ‘the best sounding board you can get.’

Watch clips from Christian Fløe Svenningsen’s presentation.

DJ Noize’s personal experience with AI

DJ Noize (Kim Sæther) demonstrated how he uses AI-assisted tools such as ElevenLabs, StableAudio and Aiva to create music and streamline music production. AI helps create demos and pitches quickly, which is a huge advantage in the music industry. He also stressed the importance of paying for samples and having clear agreements in place.

Watch DJ Noize work live with AI on Koda’s stage.

Vita Dosa’s artistic use of AI

Producer and songwriter Vita Dosa (Sandra Vitayarat Frederiksen) explained and demonstrated how she integrates AI into her creative process. Vita Dosa uses visual AI tools to create content that inspires the creation of her music. She works with AI software like ChatGPT and DALL-E to generate lyrics and visual ideas that help her explore new creative directions.

‘I have someone to bounce ideas off in the machine,’ Vita Dosa explained, adding that AI helps her hone her artistic vision.

Watch Vita Dosa work live with AI on Koda’s stage.

Future outlooks by Sofie Hvitved

Sofie Hvitved, who is a futurist and senior advisor from the Copenhagen Institute for Futures Studies, discussed potential future scenarios for AI and music. She highlighted the potential for hyper-personalised music and new ecosystems, but also the challenges posed by the digital developments. She emphasised the importance of understanding the creative potential of AI and the ethical questions it raises.

‘The artificial intelligence we have today is the worst we will ever see,’ concluded Sofie Hvitved.

See excerpts from Sofie Hvitved’s thoughts on the future.

Bio

CHRISTIAN FLØE SVENINGSEN is partner and publishing director in the Danish independent label and music publishing company GL Music Entertainment A/S. He joined the company in 2013, after a career that has included management, label A&R, live touring and tour management. His work spans creative publishing A&R work, songwriter and producer development, as well as sync regarding pitching and music supervision alike.

SOFIE HVITVED is a futurist, senior advisor and media manager at the Copenhagen Institute for Futures Studies. She is particularly interested in the field where media and technology intersect and in what the future media landscape will look like. She is currently working with the future of the Metaverse and is head of the project Future Nordic Media Landscape 2030. Sofie has a background in the media industry, including positions at DR and as partner and CEO of the media house Luksus.

NICOLAI FRANCK works as a Tech Scout in Koda. His task is to ensure that Koda is prepared for the challenges and opportunities brought by technological developments. In this regard, AI has been by far the most prominent issue since he began working at Koda on 1 May 2023. Nicolai Franck is a former tech journalist, having covered technological developments closely for more than 10 years – for example while working for the Danish broadcasting corporation, DR. He also has a background in the music industry as an artist manager.

SANDRA VITAYARAT FREDERIKSEN is a music producer, vocalist and songwriter who has previously worked on her own releases under the alias Mendoza; she now goes by Vita Dosa. She incorporates the hyperpop mentality in her work, letting digital possibilities merge with analogue workflows. Her sources of inspiration ranges from noise to trailer music, resulting in a fusion of IDM, pop and punk. Her goal is to create a sound that is honest, heartbreaking and represents a dystopian future.

DJ NOIZE is a Danish hip hop DJ who has achieved great recognition abroad due to his skills on a pair of Technics 1210s. In 1992, Noize won his first title as a regional mixing champion in his native Denmark. Then followed five titles as Danish champion until the year 2000. He won second place in the world championships in 1994 and 1995, and first place in 1996. In 1994 he won the ‘New Music Seminar DJ Battle’, in 1998 he won ‘Turntable Techniques’. DJ Noize has entered into a large number of collaborations, including with the Danish producer Sonny B aka DJ Typhoon on releases under the title ‘Nobody Beats The Beats’. He has been part of the hip-hop group Double Mouth and is one fifth of the group Majors.

Login