JAM has developed an Artificial Intelligence platform that it says can “successfully generate high-quality music tracks using professional audio loops”. The company intends to showcase an initial wave of results from the system later next month on a new website.
As incredible as it seems, JAM claim that “the AI’s neural network can now generate a new music composition in under ten seconds”. That’s slightly quicker than the six years it took Queen to write Bohemian Rhapsody. Whether the AI can compete with Queen, though, will be a different story. The JAM software boasts a catalogue of 100,000 audio loops, which are selected by the AI and arranged to create a whole track. Genres such as EDM, techno, house, hip hop, trap and synthwave can be built as directed by the user.
JAM’s flagship product is a free music-making app, sharing the same name with the company. The new AI software aims to interact with users and collaborate, to herald the beginning of what JAM are calling the ‘Era of the AI composer.’
The software has been trained with a dataset of 3.5 million music tracks, with metadata such as moods, tempo, skip rates, likes, plays, shares, user comments and more. JAM estimates it possesses nearly 500 million data points to gather information from to create its music.
Rory Kenny, CEO of JAM explains that the “AI behaves like a DNA sequencer – effortlessly arranging and generating an infinite variety of complex loop-based music compositions. We’re just at the very beginning of our development, but the results are already impressive – and it can only get better from here”.
JAM will launch its website next month to showcase artist collaborations with the AI, with commercial release set for March 2020.