Google has continued its long strides in the field of artificial intelligence with the latest development. The tech giant is reportedly developing an AI Bot that takes text prompts as input to generate short musical pieces. Not only text prompts, but the bot can also transform a whistled or hummed melody into a piece of music. Known as Music LM, the AI Bot system is being compared with DALL-E which had the ability to produce images by using text inputs.
The Details of the New System
The information about the new system was revealed in a recently published research paper. It described MusicLM as “a model generating high-fidelity music from text descriptions”. It is expected to generate music at a frequency of 24 kHz that stays the same over a few minutes. Songs can be created from well-written and neatly-depicted captions such as
“A video game’s main music has a quick tempo, a positive vibe, and a memorable electric guitar riff. The music has startling sounds like cymbal crashes or drum rolls but is repetitious and simple to recall.” The technology is not yet ready to use, however, Google has uploaded some samples to demonstrate the kind of music that can be generated.
Is AI Bot Bigger than ChatGPT?
MusicLM is not the first generative AI Bot system for music, however, it is the first to generate songs that have high fidelity and intricate compositions. Due to its unique capabilities, the new AI Bot is being directly compared with ChatGPT.
Enunciating his views on this, Keunwoo Choi, an AI scientist at Gaudio Lab tweeted,
“To me, this is surely bigger than ChatGPT. I would say that Google has almost solved the music generation.”
Several other scientists have also expressed the same possibilities in the near future. However, only time will tell what happens when MusicLM is open to use for all.