While there are numerous AI-powered music creation tools available, Adobe has unveiled its unique take on the concept.
During the Hot Pod Summit in Brooklyn, Adobe revealed Project Music GenAI Control, a platform that can produce sound from text descriptions such as “happy dance” or “sad jazz,” or from a reference melody, allowing users to customize the results within the same process.
With Project Music GenAI Control, users can modify elements like tempo, intensity, repeating patterns, and structure. They can also lengthen a track, remix music, or create an endless loop.
Developed in collaboration with researchers from the University of California and Carnegie Mellon, Project Music GenAI Control may potentially be accessible to the public in the future. However, as of now, it remains in the experimental phase, as noted by Adobe’s head of audio and video AI research, Gautham Mysore, during a panel at Hot Pod — the platform currently lacks a user interface.
“This concept emphasizes the idea that AI generates music while putting you in control as the director, offering a range of possibilities to explore. You don’t need to be a composer to express your musical ideas,” Mysore explained.
AI-generated music tools, and AI tools in general, have raised ethical and legal concerns as AI-generated music, art, and text become more widespread.
Homemade tracks that utilize GenAI to create recognizable sounds, lyrics, and vocals that can be perceived as authentic have gained popularity. Music labels have swiftly issued takedown requests citing copyright concerns. However, the issue of whether “deepfake” music infringes upon the intellectual property of artists, labels, and other rights holders, particularly when GenAI music tools are trained on copyrighted content, remains ambiguous.
A federal judge ruled in August that AI-generated art cannot be copyrighted. Nonetheless, the U.S. Copyright Office has yet to take a definitive stance, only recently initiating discussions on copyright matters concerning AI. It is also uncertain whether users might be held liable for copyright infringement if they attempt to commercialize music generated in the style of another artist.
Mysore mentioned that Adobe typically develops its GenAI tools using data that is either licensed or in the public domain to avoid potential IP conflicts. However, there is no confirmation if this approach will be applied to Project Music GenAI Control. He further stated that Adobe is working on watermarking technology to help identify audio created by Project Music GenAI Control but admitted that it is still a work in progress.
“Adobe is taking a responsible approach to these matters. There are many talented musicians producing this content… I believe they and tools like Project Music GenAI Control can coexist, leading to the emergence of new musical ideas,” Mysore concluded.
