We’ve recently seen artificial intelligence (AI) being used to help people create images and write content. And now, there’s a new AI system from Google that can generate music by following text descriptions.
The search and advertising company has named this system ‘MusicLM’. In an academic paper Published on January 26 last.Google researchers describe MusicLM as “a model that generates high-fidelity music from text descriptions such as ‘a quiet violin melody backed by a distorted guitar riff’.”
“We show that MusicLM can be conditioned on both text and a melody such that it can convert whistled and hummed melodies in a manner described in text captions,” the paper reads.
According to the research paper, MusicLM can take cues from user-generated descriptions such as “Enchanting jazz song with a memorable saxophone solo and a solo singer” or “Berlin ’90s techno with low bass and a strong kick,” and Can get relevant results.
You can listen to some of the tracks produced by MusicLM. Here.
The development of Google’s MusicLM follows the rapid growth of OpenAI’s ChatGPT, an AI-powered natural language processing tool. Chat GPT was so popular that Google’s management Reportedly announced a “Code Red”. According to a New York Times report. As a result, the company aims to More than 20 AI-related projects are set to be introduced this year.Including an AI-powered version of Google Search.
However, don’t expect MusicLM to be available to the public anytime soon. According to Google researchers, there are concerns about programming biases that could result in technical glitches, lack of representation, and “potential misuse of creative content.” In fact, during one experiment, Google researchers found that about 1% of the music MusicLM produced was copied directly from the songs it was trained on.