As AI slowly creeps it way into every facet of our digital lives — from essay writing to conversations with therapists to the generation of original art — it's the least bit surprising that artificial intelligence would also take a step into music. It's also not surprising that tech giant Google would be the first major player on the scene.
The company is reportedly building an AI bot that can create "original" music from both text and sound prompts — users would be able to type in increasingly specific prompts noting genres and styles or even build songs based on a hummed or whistled melody, according to reports. The future app is known in-house as MusicLM.
The information was revealed in a research paper released on Jan. 26, describing MusicLM as a "model generating high-fidelity music from text descriptions" that "generates music at 24 kHz that remains consistent over several minutes." The paper explains that songs can be created from richly written captions, such as:
The main soundtrack of an arcade game. It is fast-paced and upbeat, with a catchy electric guitar riff. The music is repetitive and easy to remember, but with unexpected sounds, like cymbal crashes or drum rolls.
Additional sequences of timed text prompts help build the structure of songs, along with a library of sounds and other AI prompts generated from sources like art archives.
Examples of the AI-generated songs have already been posted to Google's Github account, part of a preliminary release of a 5,500 music-text pair dataset known as MusicCaps.
The unveiling of such a platform will inevitably stir additional conversations about the role of artificial intelligence in intellectual property theft and copyright infringement, generated by a plethora of artists and art repositories who haven't consented to public use of their art in the creation of AI bots like these — others, meanwhile, are capitalizing on the surge in AI fronted tech. Additional AI developments also pose unique risks for the humans behind the tech as well, as exploited workforces face the brunt of data mining and moderation.
As for now, the Google AI music maker is not going to be released any time soon, with the company explaining ongoing concerns about cultural programming biases, glitches, and concerns about plagiarism that have to be resolved before its launch.
from Mashable https://ift.tt/S2yCtLg
https://ift.tt/sZYxQ8v
Comments
Post a Comment