Get Back

Summary

Get Back builds physical AI robots that generate music in real-time jam sessions, allowing musicians to interact and improvise through their instruments, without ever touching a computer.

What if guitar players or drummers could directly interact with creative AI software via their instruments, without having to type, prompt or click on their computer? Get Back develops intuitive interfaces between musicians and AI-driven robots to explore the creative possibilities of AI music. What if we could interact with AI during jam sessions and use body language, voice commands and musical call-and-response and to co-create?

 

The Get Back project includes a series of physical AI robots, jam sessions, recording sessions, an EP release, an AI Song Contest entry, alongside public presentations and workshops.

 

 

Challenge

Get Back tackles the disconnect between traditional musicianship and current generative AI tools. Musicians like guitarists and drummers often find themselves bound to a computer mouse, keyboard, and screen when working with AI. But what if your creative process doesn't involve prompting or clicking? And if we manage, how can we keep creative control and empower musicians, rather than replacing them?

 

 

Result

The Get Back project delivers four AI-driven robots enabling seamless collaboration with musicians through intuitive physical interfaces. Key outcomes include jam sessions with advanced AI-music interaction, personalized parameters for creative control, and diverse artistic expressions. Additionally, we document the entire process to inspire others to explore and innovate in this field, fostering a broader conversation around the intersection of music and AI.

 

 

Way of working

Thunderboom works together with experienced musicians, designers and developers to build and test the AI robots. The Get Back project employs a design-driven, iterative approach, integrating musician feedback in every stage. Using the Double Diamond model, prototypes evolve through research, definition, development, and delivery phases. Open-source principles and real-world jams refine AI's adaptability to instrumentalists. Collaborative testing ensures both artistic and technical success, culminating in inclusive, impactful outcomes.

 

 

Partners

Jaimie van Hek, Pien den Hollander, Ivo Koolen, Bas Maat, Situated Art & Design Research Group, Avans Hogeschool, Innovation in the Music Industry Research Group, Inholland Hogeschool.

 

This project is made possible with the help of Stimuleringsfonds.