“Using artificial intelligence, they recorded eight tracks in 24 hours – and you’ll be surprised how it was done.”
Hip-hop artists Nimrod Cain and Vinay Pai teamed up with an experimental AI tool to see if they could generate enough fresh beats and lay down vocals for a complete album in just one day. Their experiment yielded unexpected results and reveals the cutting-edge – yet untapped – potential of generative AI in music creation.
The AI Music Producer that functions like ChatGPT
The AI system the duo worked with, called MusicGen, operates similar to OpenAI’s ChatGPT by taking natural language prompts and generating musical compositions in return. Developed by Meta, MusicGen has been trained on over 20,000 hours of audio clips to recognize different musical styles and produce coherent 30-second snippets on demand.
Cain and Pai tested MusicGen’s abilities by feeding it prompts describing beats influenced by legends like Wu-Tang Clan as well as modern rap styles like trap and drill. To their amazement, the AI was highly skilled at generating lo-fi, drill, jazz, and classical compositions – making a wealth of raw material available for sampling or building full tracks around.
Read more: No idea how to create music via AI? Browse our AI tool database.
Recording a complete album in just 24 hours
Armed with this powerful creative tool, the pair booked cheap studio time in New York and Oakland determined to finish a full EP in one marathon session. Moving at lightning speed, they explored MusicGen’s vast sound libraries, layered vocals, and brought in collaborators to guest on tracks. Topics ranged from war, technology, religion, and capitalism recorded over trap-like beats.
———
Want to stay ahead in AI and never miss any news?
Get the hottest AI music news every week + AI tool tutorials, strategies & growth tips.
———
By dawn the next day, eight complete songs were in the can – proof that generative AI has the potential to significantly speed up the creative process and open new artistic frontiers when used respectfully by human collaborators. A documentary of the whirlwind session reveals just how transformative technologies like MusicGen may be for music-making.
Is AI a threat to music jobs?
Not everyone shares the artists’ enthusiasm, however. Some express valid concerns that generative AI could displace jobs across the music industry by supplying an endless stream of finished compositions. But like any new technology, MusicGen is still in its early stages and will likely evolve to augment human creativity rather than replace it. As long as oversight ensures AI respects creative control and credit, it has exciting potential to fuel artistic expression in new forms.
So in summary, this one-day experiment showed how AI and human collaboration can break boundaries. By sparking imagination and turbocharging output, tools like MusicGen may light the path to whole new genres of computer-assisted music. The results also raise intriguing discussions around AI’s role in 2023 and beyond.
Meta AI and 10:22pm Records developed MusicGen for research purposes. All proceeds from the “Sensory Overload” album recorded during this session were donated to Beats Rhymes & Life, an Oakland nonprofit using hip-hop for youth outreach.
