In 30 seconds
In this 2024 demonstration, musicians use Music AI Sandbox to generate and select musical material, experiment with descriptive prompts and add texture to loops.
- 1A virtual crateJean selects generated fragments as he would search records for samples.
- 2Describing a moodTranter and Blush use descriptive prompts to explore sounds.
- 3Filling out a loopRebillet tries new instrumental textures and chooses what to keep.
Takeaways
- Generated material can serve as a starting point for sampling and selection.
- Descriptive prompts can help explore a desired mood.
- Musicians retain creative choices about which outputs to use.
Read the edited text by chapter
Sampling and experimentation
Wyclef Jean experiments with Music AI Sandbox, a set of tools developed by Google DeepMind and YouTube with musicians. He tries combining musical influences and listens for unexpected results.
Wyclef Jean experiments with Music AI Sandbox, a set of tools developed by Google DeepMind and YouTube with musicians. He tries combining musical influences and listens for unexpected results.
Jean compares the process to a hip-hop producer digging through records for a sample. Instead of searching physical vinyl, he listens to generated material, selects a fragment and incorporates it into a song.
Songwriting with prompts
Songwriter Justin Tranter experiments with chords and describes the appeal of expressing ideas in a familiar musical language, even without being a specialist in modern production software.
Songwriter Justin Tranter experiments with chords and describes the appeal of expressing ideas in a familiar musical language, even without being a specialist in modern production software.
Tranter and collaborator Blush try atmospheric prompts, using places and eras to suggest a mood. They listen to the outputs and adjust their direction as they work.
Adding texture to loops
Marc Rebillet uses generated elements to fill gaps in a loop, experimenting with percussion, viola and clapping. He describes the tool as a collaborator offering ideas he can accept, reject or reshape.
Marc Rebillet uses generated elements to fill gaps in a loop, experimenting with percussion, viola and clapping. He describes the tool as a collaborator offering ideas he can accept, reject or reshape.
Jean says the tools can speed up moving ideas from his head into a track. The video shows demonstrations by participating artists; their enthusiasm is their experience, not a guarantee of a particular result for every musician.
More from this channel
38:10EN
Google DeepMind· Oct 1, 2026From deepfakes to DNA: how SynthID watermarks AIHannah Fry, Pushmeet Kohli, Jeremy Ratcliffe · Google DeepMindHannah Fry sits down with Pushmeet Kohli and Jeremy Ratcliffe to explain SynthID: a mathematical watermark now used on text, images and video, and being extended to protein sequences so DNA synthesis companies can tell AI-designed molecules from nature.
44:25ENGoogle DeepMind· Sep 9, 2026How AI Is Changing Weather ForecastingHannah Fry, Peter Battaglia · Google DeepMindHannah Fry and Peter Battaglia discuss hurricane forecasts, uncertainty and WeatherNext 3.
44:41ENGoogle DeepMind· Aug 26, 2026The Mathematics of AI UncertaintyZoubin GhahramaniZoubin Ghahramani explains Bayesian updating, calibration and their computational challenges.
10:52PT → EnglishGoogle DeepMind· Jul 14, 2026Reconstructing Pelé’s Lost GoalHistorical research, actors and AI recreate a 1959 play without original footage.
53:05ENGoogle DeepMind· Jul 10, 2026Inside AI Models: Neel Nanda on Interpretability and SafetyHannah Fry, Neel Nanda · Google DeepMindHannah Fry explores reasoning traces, probes and sparse autoencoders.
42:38ENGoogle DeepMind· Jun 23, 2026When AI Agents Delegate: Security and a Possible Agent EconomyHannah Fry, Nenad Tomašev · Google DeepMindNenad Tomašev discusses specialists, oversight and correlated failures with Hannah Fry.