About this episode
Why Serious Composers Should Not Use AI as a Creative ToolThere is a growing pressure on composers to embrace AI as a creative partner—an insistence that resisting it is nostalgic, elitist, or fearful of progress. This framing is false. Serious composers are not rejecting technology; they are rejecting the misrepresentation of authorship.AI may be useful in processes.It can never be justified as a tool for creative composition.And the difference matters.Creation Is Not AssemblyComposition is not the act of assembling pleasing patterns.It is the act of choosing—under constraint, risk, memory, failure, and intention.Every meaningful composition is the result of:lived experienceaesthetic judgementphysical interaction with soundcultural placementemotional consequenceAI possesses none of these.It does not intend.It does not hesitate.It does not risk being wrong.It predicts.AI Is Fundamentally UnoriginalAI systems do not create new musical language. They interpolate existing ones.All generative music models are trained on existing human-made work, statistically analysing:pitch relationshipsrhythmic tendenciesharmonic probabilitiesstylistic signaturesWhat emerges is not originality, but averaged familiarity.This is not an insult—it is a technical fact.AI cannot:reject precedentbreak a system it does not understanddevelop a personal syntaxrespond to silence as meaningcreate tension through restraintinvent form through failureEvery “new” result is a recombination of what already exists.Groundbreaking art, by contrast, often fails before it works.AI never fails. It only optimises.Emotion Cannot Be Simulated Into ExistenceMusic does not contain emotion.Emotion emerges through human perception of intentional gesture.When a composer distorts time, fractures form, or denies resolution, the listener senses a human struggle behind the sound.AI has no inner life to encode.No body to resist.No fear of exposure.No personal stake.An AI-generated lament is not sad.It merely resembles music that once accompanied sadness.This distinction is not philosophical—it is perceptual. Listeners intuitively detect when music lacks human risk.Tools Are Not the Same as AuthorsThere is an important and often deliberately blurred distinction here.AI-like systems have been used in music for decades in non-creative roles, including:pitch correction and tuning analysistempo detection and alignmentaudio restoration and noise reductionorchestration mock-upsscore layout and notation optimisationspectral analysis and timbral visualisationrecommendation systemsadaptive mixing and mastering assistanceThese tools operate after or around human creative decisions.They do not decide:
