13:00 - 13:05: Introduction by moderator Marie Høst
13:05 - 13:50: Double keynote on AI and music: Art machine or Copy machine? When AI remembers too much by Anders Søgaard, Professor of AI at the University of Copenhagen, and Nicolai Franck, Tech Scout at Koda
Music creation has always been the domain of humans – until now. Generative AI can produce and imitate existing music at a speed that challenges artists, rights holders, music users, lawmakers, and our culture.?In this double keynote, we focus on the clash between generative AI and copyright, diving into why AI models sometimes create pure plagiarism.
Nicolai Franck, Tech Scout at Koda, will first give a quick overview of the lawsuits against AI music companies that have gained momentum over the past year and the paradox that AI companies acknowledge that their technology cannot function without access to protected works, but they do not believe they should pay for it.?With concrete examples, he will also demonstrate how AI music services plagiarize megahits and specify how AI, without regulation, threatens the income of human artists and a sustainable economy in the music industry.
Next, Anders Søgaard, professor in AI and language models at the University of Copenhagen, will explain how a generative music AI works. What does it mean for an AI to "remember" parts of its training material – and is that a flaw or a feature? Where is the line between inspiration and copying when a machine learns from millions of songs? And is it even possible to identify which songs are included in the output that an AI generates?
Speakers: Anders Søgaard, professor in AI, University of Copenhagen // Nicolai Franck, Tech Scout, Koda
Moderator: Marie Høst, journalist og moderator
This session is part of the theme block Let's Talk AI & Music.
The session will be conducted in Danish.
Presented by Koda in collaboration with Promus and SPOT+.
