Generative AI and the future of music streaming

Refreshments will be available from 3.30pm. Please note that this event takes place in the Club Room, Churchill College, not the CMS Computer Room.

Generative AI has emerged as one of the most disruptive forces in the music industry in recent years. For better or worse, it is fundamentally reshaping how music is created, distributed, and consumed at scale. In this talk, we explore its impact on Deezer, an international music streaming platform.

Over recent years, Deezer has witnessed a dramatic surge in AI-generated content, to the point where artificially produced tracks now outnumber those created by human artists. In response, we have developed a series of initiatives aimed at detecting and mitigating the presence of such content across our catalog, with the dual objective of preserving the integrity of our ecosystem and protecting real artists' interests.

Central to this presentation is a detection system designed to automatically classify tracks as either AI-generated or human-produced. We describe the technical foundations of our approach and discuss the challenges inherent in building robust, scalable detection within a live, large-scale streaming environment.

Biographies

Gabriel Meseguer-Brocal is a senior research scientist at Deezer with over 10 years of experience in the music industry. Before joining Deezer, he completed postdoctoral research at Centre National de la Recherche Scientifique (CNRS) in France. In 2020, he earned his Ph.D. in Computer Science, Telecommunications, and Electronics with a focus on the Sciences & Technologies of Music and Sound at IRCAM. His research interests include signal processing and deep learning techniques for music processing, with a focus on areas such as source separation, self-supervised learning, dataset creation, AI-detection, and multimodal analysis.

Bruno Sguerra is a senior research scientist at Deezer. His research focuses on understanding user behaviour in music streaming environments, with particular interest in patterns of repetition, discovery, search, and contextual listening. He develops user models for recommendation systems, drawing on behavioural and psychological frameworks such as the mere exposure effect and the modelling of uncertainty in user feedback. Bruno has been actively involved in the research community, serving as a meta-reviewer for The Web Conference (WWW) and as a reviewer for RecSys. He co-organised the RecSys 2026 Challenge on conversational recommender systems.

Yuexuan Kong was a PhD student at Deezer in collaboration with Centrale Nantes. Her research focused on self-supervised learning and representation learning for music information retrieval. She worked on designing musically informed loss functions and developing efficient training methods for compact models capable of encoding rich musical information in their representations.

Also available on Zoom: https://cam-ac-uk.zoom.us/j/82101982541?pwd=O91QJsFJXpJubeafEvqPp6xNou1rbV.1