Artificial intelligence is transforming the way music is created, distributed, and experienced. From cleaning up century-old recordings to spawning “deepfake” hits, AI tools are carving new paths and raising fresh challenges for artists, producers, and fans alike.
AI as a Creative Assistant
One of the most celebrated uses of AI today is stem separation. This refers to the ability to isolate vocals, drums, bass, and other elements from an existing track. This technology enabled Sir Paul McCartney to resurrect a decades-old John Lennon vocal and complete what fans now know as the final Beatles song.
Importantly, AI wasn’t used to generate Lennon’s voice; it simply cleaned and separated the original recordings so McCartney could craft a polished, authentic duet. Tools like stem separation make restoration projects smoother, whether the goal is remixing, mastering, or archival preservation.
Looking ahead, stem separation could revolutionise karaoke apps that let users mute lead vocals in real time, educational platforms that teach mixing by exposing individual stems, and rights management systems that automatically flag unlicensed sample usage. As these models improve, the barrier to entry for sophisticated audio editing will shrink further, empowering bedroom producers and heritage-label archivists alike.
The Dark Side of Deepfakes
However, the very same algorithms that clean and isolate can also impersonate. In early 2023, an AI-generated “Fake Drake” single went viral, fooling listeners into believing the rapper himself had dropped an unauthorized track.
The clip spread quickly on social media before platforms and labels scrambled to pull it down. This incident highlights how easy it is to fabricate plausibly “real” music without any artist involvement and without their express permission. As generative models get more convincing, labels and law firms will face an uphill fight trying to control unlicensed or misleading content.
Jobs on the Line?
Beyond copyright headaches, there’s a human cost. AI can draft melodies, craft lyrics, and even suggest chord progressions, the tasks traditionally reserved for session musicians, arrangers, and songwriters. A recent analysis asked whether AI will one day “take overt” musician jobs; the worry is palpable among gigging artists already juggling side-hustles to make ends meet.
In the UK, the first Musicians’ Census found that although 70% of professionals hold degrees (half in music), the average annual income from “music work” is just £20,700, with 43% earning under £14,000 and reliant on additional employment elsewhere. With AI tools poised to automate parts of composition and arrangement, musicians who supplement their craft with teaching, live-gigging, or session work may find it harder to compete and make a living from music, in general.
Striking the Right Chord
AI in music isn’t inherently villainous. When used thoughtfully, it preserves our sonic heritage, expands creative possibilities, and helps indie artists punch above their weight. The challenge lies in regulating misuse, protecting artist rights, and ensuring that technological gains don’t hollow out the very community that gives music its heart. Balancing innovation with integrity will determine whether AI becomes a true collaborator or an unwelcome replacement in the industry’s next chapter.
