We’ve recently seen artificial intelligence (AI) being used to generate images and help people write content. And now, there’s a new AI system from Google that can generate music by following a text description.
The search and advertising giant named this system ‘MusicLM.’ In an academic paper published last January 26, Google researchers describes MusicLM as “a model generating high-fidelity music from text descriptions such as ‘a calming violin melody backed by a distorted guitar riff’.”
“We demonstrate that MusicLM can be conditioned on both text and a melody in that it can transform whistled and hummed melodies according to the style described in a text caption,” the paper reads.
According to the research paper, MusicLM can take cues from user-generated descriptions like “enchanting jazz song with a memorable saxophone solo and a solo singer” or “Berlin 90s techno with a low bass and strong kick,” and receive corresponding results.
You can listen to some of the tracks generated by MusicLM here.
Google’s development of MusicLM comes after the rapid rise of OpenAI’s ChatGPT, a natural language processing tool powered by AI. ChatGPT was so popular that Google’s management reportedly declared a “code red,” according to a report by The New York Times. As a result, the company is purportedly set to introduce over 20 AI-related projects this year, including an AI-powered version of Google Search.
MusicLM isn’t the first AI music generator. As TechCrunch points out, Riffusion, Dance Diffusion, Google’s own AudioLM, and OpenAI’s Jukebox have experimented on the subject.
Don’t expect MusicLM to be available to the public in the near future, however. According to Google’s researchers, there are concerns about programming biases that may result in technological glitches, lack of representation, and the “potential misappropriation of creative content.” In fact, during an experiment, Google’s researchers found that about 1% of the music that MusicLM generated was directly copied from the songs it was trained on.