M5: The JAM_BOT: An Insight Into An AI-Human Co-Created Musical Improvisation

Perry Naseck, [email protected], Research Assistant, MIT Media Lab. Perry Naseck is an artist, engineer, and researcher working with interactive, kinetic, light- and time-based media. He specializes in interaction, orchestration, and animation of systems of sensors and actuators for live performance and permanent installation. Perry is pursuing a PhD in the Responsive Environments group at the MIT Media Lab where he focuses on creating new experiences for musicians and audiences. His research explores how to closely link live performance visual systems with a focus on improvised and AI-generated music. Perry completed a Bachelor's of Engineering Studies and Arts at Carnegie Mellon where he studied both Art and Electrical & Computer Engineering.

Lancelot Blanchard, [email protected], Research Assistant, MIT Media Lab. Lancelot Blanchard is a musician, engineer, and AI researcher pursuing a PhD at MIT Media Lab's Responsive Environments group. His research explores the development of generative AI systems that can enhance the creative processes of musicians, with a special focus on live performances. Lancelot has collaborated with multiple GRAMMY-winning artists to create innovative tools for real-time AI-augmented music making. Lancelot holds a Master's of Research in Artificial Intelligence and Machine Learning and a Master's of Engineering in Computing from Imperial College London, as well as a degree in classical piano from the Conservatory of Rennes, France.

This show is a demonstration of the JAM_BOT, a real-time system for collaborative free improvisation with music language models that was developed as part of a collaboration between MIT Media Lab researchers and GRAMMY-winning visiting artist Jordan Rudess. The JAM_BOT accompanies the improvisation of live performers onstage using optimized Transformer-based Music Language Models. It was first demonstrated publicly during a sold-out performance held at the MIT Media Lab on September 21, 2024. This ISMIR 2025 show intends to reproduce segments of the original show to demonstrate the potential of the system to the ISMIR scientific community.