We Corrected the Humanity Out of Music

Modern music has not run out of talented artists. There are probably more technically capable singers, producers and musicians recording today than at any previous point in history. Equipment has become better and more accessible, recording software can run on an ordinary laptop and somebody can create an entire album from a bedroom without needing an expensive studio, a major record label or permission from an executive who thinks every unusual creative decision is commercially dangerous.

The problem is what happens to that talent after it enters the modern music industry. An artist may begin with a recognisable voice, strange influences, unusual ideas and something they genuinely want to communicate, but those qualities are then passed through an enormous production system designed around consistency, commercial safety and immediate attention. Pitch correction cleans the vocals, editing removes the awkward breaths, streaming data influences the structure, TikTok requires a section that can survive as a short clip, Instagram requires a marketable identity, brands require somebody safe enough to advertise beside and executives need evidence that the entire thing will generate money.

None of those individual components is necessarily bad. Auto-Tune can rescue an incredible vocal performance containing one imperfect note. Editing can remove accidental noise without destroying the emotion of a recording. TikTok can expose an unknown musician to millions of people, while streaming services allow nearly anyone to release music without persuading a record shop to stock it. Brand deals can provide artists with enough money to continue creating, and professional producers can turn a rough idea into something far better than the artist could have produced alone.

The problem begins when every part of that system is connected and pointed towards the same objective. The artist is no longer simply making a song and using technology to express it properly. The song is being adjusted at every stage to become easier to market, easier to recognise, easier to place inside a playlist and easier to extract into the short section that somebody might hear while scrolling through their phone.

The result is not always bad music. In many cases, it is extremely competent music. The vocals are clean, the instruments are perfectly positioned, the chorus arrives at the expected moment and the entire recording sounds as though nothing has been left to chance. You can dance to it, place it inside an advert, use it behind a TikTok video and listen to it repeatedly without encountering anything particularly uncomfortable.

That consistency is precisely the problem. A factory does not necessarily produce bad products. It produces predictable products, and the modern music industry has become exceptionally good at taking talented people and making them predictable.

Auto-Tune Was Only One Part of the Factory

Auto-Tune is normally blamed whenever people complain about modern music, but Auto-Tune by itself did not destroy anything. It was introduced as a tool for correcting pitch, and there are obvious situations where that is useful. A singer may deliver an emotionally perfect performance while missing one note. Throwing away the entire take because of that one mistake would be ridiculous when software can correct it without changing everything else.

Auto-Tune can also be used deliberately as an instrument. Some artists do not attempt to hide it at all. They use the artificial sound as part of the music, in the same way another musician might use distortion, reverb or a synthesiser. Technology has always changed music, and many of the artists I love relied heavily on electronic equipment, sampling and studio manipulation.

The Chemical Brothers are not an argument against technology. Thievery Corporation are not an argument against technology. Massive Attack are not an argument against technology. These artists used machines constantly, but they used machines to create something stranger, darker and more recognisable. The technology served the identity of the music.

The modern factory frequently works in the opposite direction. The identity of the artist is made to serve the technology, the platform and the commercial structure surrounding it. Auto-Tune becomes one part of a much wider process involving vocal comping, timing correction, compression, quantisation and editing. A singer can record the same line dozens of times, after which the cleanest fragments are stitched together into one technically perfect performance that never actually happened.

Breaths can be removed, softened or moved. Words can be aligned exactly with the beat. Notes can be corrected until every transition is smooth. The volume of every line can be made consistent. The singer can sound almost identical from the beginning of the song to the end, regardless of whether the original performance became more tired, strained, emotional or unstable.

Again, none of those techniques is automatically wrong. The problem is their cumulative effect. A voice stops sounding like a person moving through a song and begins sounding like a polished surface placed on top of it.

Older recordings often contain details that would now be treated as defects. A breath arrives slightly too early. One line sounds raspier than the last. A word disappears under the instrumental. The singer sounds more tired by the end than they did at the beginning. Their voice changes because a real throat cannot produce every note with identical pressure, tone and texture.

Those details remind you that somebody stood in front of a microphone and physically performed the song. The voice was not simply a clean information channel carrying lyrics. It was the sound of a body attempting to express something.

Modern production has become very good at removing the evidence that the body was ever there.

The Music Industry Is Optimising for Platforms

The modern music factory does not end inside the studio. The finished song must then survive platforms built around attention.

TikTok and Instagram did not invent short songs, catchy hooks or repetitive choruses. Musicians have always wanted people to remember their work. The difference is that platforms can now measure which seconds of a song attract attention, which sections are replayed, where people stop listening and which fragments become attached to trends.

That information begins influencing what gets made. The introduction cannot take too long because the listener may swipe away. The chorus needs to arrive quickly. A section must be recognisable within seconds because it may be heard underneath somebody speaking, dancing or advertising a product. The song should contain a line that works as a caption, a reaction or a short emotional statement that can be detached from everything around it.

The artist is therefore not only writing for somebody sitting down and listening to a complete song. They are writing for somebody hearing twelve seconds through a phone speaker while looking at something else.

This does not mean music discovered through TikTok is automatically bad. A brilliant song can become popular through a short clip, and older music has repeatedly found new audiences through the platform. The problem is that the platform begins influencing the architecture of the song before the song exists.

The same thing happens with streaming. A song is no longer competing only with other songs on an album or other releases in a shop. It is competing with almost every recording ever made, all available immediately, while the listener can abandon it with one movement of their thumb.

This creates enormous pressure to become understandable quickly. The emotion must be obvious. The hook must be familiar. The production must immediately resemble something the listener already likes. Anything requiring patience becomes a commercial risk.

The industry then looks at the resulting data and concludes that people only want immediate, familiar music, despite having spent years constructing a system that constantly promotes immediate, familiar music.

People hear polished and predictable songs because platforms reward them. Platforms reward them because people respond to what they already recognise. Labels then use that response as evidence that producing anything stranger would be irresponsible.

The system trains the audience and then uses the trained audience to justify the system.

Modern Music Is Designed to Be Commercially Legible

Billie Eilish, Olivia Rodrigo, Sabrina Carpenter and Taylor Swift do not literally sound identical. Their voices are different, their production styles are different and they have moved through different genres. Billie Eilish in particular has a recognisable vocal style, while Olivia Rodrigo has drawn heavily from pop-punk and alternative rock.

It would be lazy to claim that every modern singer is talentless or that every song released by a major artist is interchangeable. Many of those artists are clearly talented, and many people form genuine emotional attachments to their music.

The more interesting problem is that even different artists are increasingly channelled into the same commercial architecture.

The song must fit playlists. The artist must fit branding. The emotion must fit social media. The release must fit a promotional schedule. The personality surrounding the music must remain recognisable enough to attract fans while being safe enough for advertisers, interviews and corporate partnerships.

A modern star does not simply release music. They operate as a complete commercial identity. Their clothes, relationships, interviews, public image, social media posts, controversies and brand deals become part of the product surrounding the songs.

That does not mean the artist is fake. It means the system has far more opportunities to interfere with how the artist is presented and what kinds of risks they are allowed to take.

A song can be catchy, technically excellent and enjoyable without being memorable in the deeper sense. You may remember the chorus because you heard it hundreds of times, but not the album it came from, the atmosphere surrounding it or the feeling of discovering something that did not resemble everything else.

A lot of modern music performs its commercial function extremely well. It becomes unavoidable for several months, appears in adverts, spreads across social platforms, enters playlists and then gradually disappears when another song is produced for the same spaces.

The factory does not always produce terrible songs. It produces songs designed to be consumed immediately and replaced efficiently.

Older Music Was Not Automatically Better

There was an enormous amount of terrible music released in the 1980s, 1990s and 2000s. Record labels manufactured artists, chased trends and released formulaic rubbish long before TikTok existed. Auto-Tune did not invent commercial music, and nostalgia can make people remember the greatest songs from previous decades while forgetting everything that deserved to disappear.

Older artists were also influenced by money, radio, labels, fashion and public expectations. They wanted people to buy records. They wanted successful singles. They were not creating music in some completely pure world where nobody cared about commercial success.

The distinction is therefore not simply old music against new music. The distinction is between music that allows technology and commercial success to support an artistic identity, and music in which artistic identity is gradually reorganised around the demands of the commercial system.

Some modern artists still produce strange, difficult and emotionally ambiguous work. Some older artists produced empty music designed entirely around sales. The difference is that the modern system has become far more efficient at measuring attention and far more capable of correcting anything considered commercially inconvenient.

It can identify which parts of a song people replay, which sections they skip, which lyrics become captions, which sounds fit current trends and which identities attract particular audiences. The factory has more information than ever before, and that information can help make successful music.

It can also make music increasingly afraid of being misunderstood.

The Chemical Brothers Understood Contrast

The Chemical Brothers are a useful example because their work proves that heavily electronic production does not need to remove humanity.

“Alive Alone,” featuring Beth Orton, appeared on Exit Planet Dust. It is not simply an electronic track with a random singer placed on top. Beth Orton’s voice becomes part of the emotional structure of the song. She sounds distant, fragile and surrounded by machinery that feels lonely rather than aggressive.

The song works because the production adapts to the vocalist. It does not feel like The Chemical Brothers created a standard template and then inserted whichever singer was available.

“Asleep from Day,” featuring Hope Sandoval, works in a completely different way. The surrounding album contains large, energetic and highly physical tracks, but “Asleep from Day” lowers the intensity. It is slow, hazy and almost suspended.

That contrast matters because an album cannot remain at maximum energy indefinitely. If every track is designed to become the loudest and most immediately exciting moment, the album becomes exhausting. Intensity only works when it is surrounded by restraint.

Hope Sandoval was the perfect voice for that role because restraint is one of the qualities that makes her recognisable. Her singing does not attempt to dominate the recording. It sounds distant, quiet and almost reluctant to occupy the centre of the song.

A modern production system might consider those qualities weaknesses. It might ask for clearer pronunciation, more projection, a larger chorus or a more obvious emotional climax. Removing those qualities would remove the entire reason the performance works.

The Chemical Brothers understood that working with a vocalist should change the music. Beth Orton did not produce the same result as Hope Sandoval. Noel Gallagher did not produce the same result as either of them. Each collaboration created a different environment around the person singing.

That sounds obvious, but it is increasingly uncommon. Modern production often changes the singer while leaving the commercial architecture underneath largely intact.

Mazzy Star Left Space for the Listener

Hope Sandoval’s work with Mazzy Star demonstrates another quality that modern commercial music often struggles to preserve: ambiguity.

“Fade Into You” does not provide a complete story. It gives you fragments involving hands, breath, shadows, blindness, darkness and emotional distance. You understand the shape of the feeling without being told exactly what happened.

The song can sound like a relationship in which one person is physically present but emotionally unreachable. It can sound like somebody disappearing into another person. It can sound like longing, grief, obsession or the gradual loss of somebody who was never completely available.

The song does not explain which interpretation is correct. That uncertainty is not a defect in the writing. It is the design.

The singer gives you emotional evidence and allows you to reconstruct what you think happened. You listen again and notice another line. You connect it to something else. You begin pattern-matching the fragments against your own memories and experiences.

You will probably never arrive anywhere near the exact event or feeling that produced the song. That does not matter. The distance between the artist’s experience and your interpretation is part of what makes the song work.

The listener becomes involved in completing it.

A completely explicit song can still be powerful, but once every part of its meaning has been explained, there may be less left to discover. A song like “Fade Into You” changes depending on the person listening and the stage of life they are in.

At one point, it may remind somebody of a relationship. Years later, the same song may remind them of a friend they lost contact with, a grandparent who died, a childhood pet, an old home or a version of themselves they can no longer return to.

The recording stays the same. The listener changes.

That is why certain songs can follow somebody for decades. They are not attached to one event. They contain enough empty space to absorb whatever loss the listener brings to them.

Great Songs Allow You to Fill in the Blanks

The best songs do not always tell you exactly what happened. They leave evidence behind.

A hand, a breath, a memory, a colour, a shadow, a physical sensation or a sentence that appears disconnected from everything around it can communicate more than a complete description of an argument.

The listener takes those fragments and builds a story. They try to understand what the singer has been through, who they are singing about and why certain images were chosen. The listener knows they may be wrong, but the attempt becomes part of the relationship with the song.

This is similar to looking at an incomplete painting. The artist provides the outline, the atmosphere and some of the colours, but deliberately leaves sections unresolved. Every person looking at it fills those sections differently.

Modern commercial music is often more direct because directness is easier to distribute. A lyric about a breakup can be immediately understood, extracted and shared. The listener knows who left, who is angry, who regrets it and which sentence should become the caption.

There is nothing inherently wrong with writing directly about a boyfriend, girlfriend or breakup. Musicians have written about relationships for as long as music has existed. The problem appears when the emotional structure becomes so predictable that the listener can identify every stage before the song reaches it.

The boyfriend left. The singer is hurt. The singer becomes angry. The chorus announces that they are now stronger. The bridge contains the emotional realisation. The song ends with closure.

Real memories do not normally behave like that. Grief does not arrive as a clean story. It appears as a smell, a place, a fragment of conversation, the weather on a particular day or a memory that returns without permission.

The best dark music understands that emotional confusion is often more truthful than emotional explanation.

Dark Music Gives Your Feelings Somewhere to Exist

A dark song does not need to describe the exact event that happened to you. It only needs to leave enough room for you to place your experience inside it.

Somebody can hear “Fade Into You” and think of a failed relationship. Another person can hear it and think of a friend who died. Somebody else can connect it to a pet, a grandparent, a traumatic memory or a period of their life they still cannot completely escape.

The original meaning of the song matters, but it is not the only meaning the song is allowed to have.

This is why dark and ambiguous music can feel strangely comforting. It does not necessarily improve your mood or tell you that everything will be fine. It does something more useful. It gives a difficult feeling a shape.

The song becomes evidence that another person has experienced something resembling the loneliness, desperation or grief you feel, even if their circumstances were completely different. It makes the feeling seem less isolated.

People do not always listen to sad music because they want to become sadder. Sometimes they listen because happiness feels dishonest at that moment. A dark song does not demand that they recover immediately. It sits beside the emotion and allows it to exist.

That is an incredibly powerful function of music, and it is difficult to measure. A streaming platform can record whether somebody replayed the chorus. It cannot properly record why one line reminded them of somebody they had not thought about in ten years.

A dashboard can measure engagement, but it cannot measure emotional recognition. A great song does not simply tell you that somebody was hurt. It gives your own hurt somewhere to sit.

Beautiful Drug Turns a Person Into an Addiction

Thievery Corporation’s “Beautiful Drug” demonstrates how lyrics can suggest a story without reducing it to a literal explanation.

The song uses the language of addiction, but the emotional meaning can be understood as becoming addicted to another person. Memories return like withdrawal. The person remains present inside the singer even after the relationship has ended. The desire is not simply to remember them but to experience the feeling again, despite knowing that returning may be destructive.

The person becomes the drug moving through the singer’s veins.

That interpretation is not delivered as a clean statement explaining that a relationship ended and the singer still misses the other person. The meaning is built through images of dependence, memory and physical craving.

You can listen to the song without studying any of that. The production and vocals work without requiring an analysis of every line. But if you eventually begin listening more closely, another layer appears underneath what you already enjoyed.

That layering gives the music depth. The song can operate as sound, atmosphere, emotion and narrative at the same time.

Many people do not begin by analysing lyrics. I normally listen to how the entire song feels before I properly process what every line means. The voice, instrumental, rhythm and atmosphere reach me first.

The lyrics become more powerful later because they explain something I was already feeling without understanding.

Modern commercial music can still do this, but the pressure towards immediate clarity makes it less common. A song designed around a short extract may need its meaning to survive outside the rest of the recording. The line needs to explain itself quickly enough that somebody can use it without hearing what comes before or after.

“Beautiful Drug” rewards listening to the complete song. The factory rewards extracting the most marketable sentence.

Thievery Corporation Created Music for Themselves First

Thievery Corporation are one of the strongest examples because their music is heavily produced, electronic and often built from multiple cultures, genres and vocal styles. Their work cannot be dismissed as technically raw or opposed to modern production.

The authenticity comes from artistic intention rather than primitive recording.

Radio Retaliation, released in 2008, moves through Brazilian influences, jazz, reggae, electronic music, political themes and songs that barely seem to belong to the same conventional genre. Tracks such as “Beautiful Drug,” “33 Degree” and “La Femme Parallel” do not sound as though they were created by beginning with a market category and asking how to satisfy it.

They sound like the result of artists deciding what they wanted to explore and trusting that somebody might follow them.

The album can move between languages, cultures and political ideas without constantly explaining why each decision is commercially sensible. Some tracks feel confrontational. Others feel detached, romantic or almost weightless.

It is the type of music that can make you ask what the fuck you have just listened to and then immediately play it again.

That is not because every song is perfect. It is because the album has an authorial identity. Somebody made decisions that were not primarily designed to resemble the most successful song released the previous month.

It was obviously still made for people to hear. Thievery Corporation wanted an audience, sold records and performed concerts. The distinction is that the imagined audience does not appear to have dictated every creative decision before it was made.

There is a difference between making something for an audience and allowing an audience to determine what you are permitted to make.

The modern music industry increasingly confuses those two things.

Massive Attack Made Darkness Physical

Massive Attack understood how production, lyrics and atmosphere could combine into something that felt dark before the listener had consciously interpreted the words.

“Teardrop” is one of the clearest examples. The lyrics do not describe a complete conventional story. They are bodily, fragmented and abstract. Love is treated as an action rather than a simple emotion. Breath, feathers, movement and physical instability create a feeling of vulnerability without explaining exactly what happened.

The words sound as though they came from the emotional aftermath of something rather than from somebody calmly documenting the event.

The production makes those fragments more powerful. The rhythm feels almost like a heartbeat. Elizabeth Fraser’s voice sounds delicate but uneasy. The song is beautiful, but it never feels completely safe.

That is something Massive Attack repeatedly achieved. Their music did not simply discuss darkness. It made darkness physical. The bass, space, repetition and vocals created anxiety before the lyrics had been understood.

This proves again that polished production is not the enemy. Massive Attack were not leaving every recording untouched. Their music was carefully constructed, processed and engineered.

The difference is what the engineering was attempting to accomplish. The production did not remove ambiguity. It intensified it. It did not clean the darkness until it became commercially comfortable. It made the listener remain inside it.

Everything but the Girl Sounded Like Themselves

Everything but the Girl are another example of artists whose identity remained recognisable even as their sound changed.

Their earlier work and the electronic direction of Walking Wounded do not sound identical, but Tracey Thorn’s voice continues to carry a particular restraint and emotional distance. The production can become cleaner and more electronic without turning her into a generic vocalist.

That is what artistic evolution should look like. The sound changes, but the person does not disappear.

Modern commercial music often treats consistency and recognisability as branding requirements, but that is different from identity. Branding means repeating the surface features that audiences already associate with an artist. Identity means retaining the deeper way the artist interprets emotion, sound and language even when the surrounding music changes.

An artist with a real identity can make two completely different records and still sound unmistakably like themselves.

A manufactured identity can repeat the same style for years and still feel as though nobody is actually there.

Tame Impala Proves Modern Music Can Still Be Strange

Tame Impala is one of the strongest counterarguments to the idea that modern music has become universally empty because Kevin Parker is a completely modern artist. His music is heavily produced, electronically manipulated and filled with effects. His vocals are layered, compressed, distorted and surrounded by synthesised sounds. He is not attempting to recreate the recording conditions of Mazzy Star or Massive Attack, and his success has occurred during the same streaming and social-media era that produced the factory I have been describing.

The difference is that Tame Impala still sounds unmistakably like Tame Impala.

Kevin Parker has a visual identity, a musical identity and a completely recognisable way of constructing sound. You can normally hear a Tame Impala song for several seconds and understand what you are listening to before anybody tells you the artist. The swirling effects, unusual drums, enormous synthesised textures, distorted guitars and vocals that sound as though they are floating somewhere between the singer and the instrumental all belong to the same world.

He is not necessarily the greatest conventional singer in the world, but that is part of why the music works. His voice does not need to demonstrate perfect technical ability because it has character. It can sound thin, distant, fragile, strained or completely buried underneath the production, and those qualities become part of the identity rather than problems that need to be corrected.

The music is slightly mad, and people love it because it is slightly mad. It does not sound as though a committee took an ordinary pop song and added psychedelic effects after deciding that psychedelic music was currently popular. It sounds like somebody has a genuinely unusual way of hearing music and has been given enough control to make everybody else hear it too.

Tame Impala also has a complete visual identity. The album artwork, live performances, psychedelic colours and surrounding imagery feel connected to the sound. It is not simply a collection of singles packaged underneath a photograph of whichever artist is currently marketable. Each album attempts to create a complete world.

That world can still become commercially successful. Tame Impala songs appear in films, adverts, playlists and social-media videos. Kevin Parker has collaborated with enormous mainstream artists, and his music has generated the same kind of commercial attention that supposedly requires musicians to become predictable.

The success does not disprove the identity. It demonstrates that identity can be commercially successful when somebody is allowed to retain it.

Tame Impala proves that modern production is not the problem. Technology is not the problem. Layered vocals, electronic instruments, effects and pitch correction are not automatically the problem. Kevin Parker uses modern technology to push his instincts further rather than gradually removing every recognisable quality until the music can sit beside anything else without causing disruption.

The factory metaphor does not apply equally to every modern musician. The problem is not being famous, using software or producing music cleanly. The problem is allowing the commercial system to become the author.

Kevin Parker still sounds like the author.

Albums Were Designed as Complete Experiences

The decline of the album as the main unit of listening has also changed how music is structured.

An album allows songs to perform different roles. One track can begin with enormous energy. Another can slow everything down. A strange instrumental can exist between two more accessible songs. A long introduction can create anticipation because the listener has already committed to the wider experience.

The Chemical Brothers could place “Asleep from Day” inside an energetic album because the quietness created contrast. The song did not need to compete as an isolated unit attempting to become the loudest moment in every playlist.

An album also creates a visual and emotional world. The cover art, sequencing, title, production and transitions become attached to one another. When somebody remembers “Fade Into You,” they may also remember the blurred purple cover of So Tonight That I Might See, the atmosphere of the surrounding songs and the period of life in which they discovered it.

The song becomes part of a complete object.

Streaming often breaks that object into individual units. A song appears between unrelated artists inside a playlist selected by an algorithm. The listener may never see the album cover for more than a few seconds, hear the next track or understand why the song was placed where it was.

This can help people discover music, but it can also remove context. The artist is no longer constructing the journey. The platform is.

The song must therefore survive without the atmosphere surrounding it. It must explain itself immediately and compete alone.

The result is music increasingly designed as a collection of independent commercial units rather than as part of a complete artistic experience.

The Difference Is Similar to SaaS and Indie Software

The difference between modern commercial music and more opinionated music is similar to the difference between enterprise SaaS software and a product created by an obsessive independent developer.

Enterprise software is normally clean, scalable, tested and designed to avoid confusing or offending anyone. Every interaction has been reviewed. Every colour has been approved. Every unusual feature requires a business case. The product is built for the broadest possible collection of customers.

It may work extremely well, but after a while every dashboard begins to look as though it was designed by the same nervous committee.

An independent developer can build something much stranger. They may introduce an interaction nobody requested, organise the entire product around their own preferences or make a decision that would never survive an enterprise planning meeting.

The result may contain rough edges. It may be less accessible to everybody. It may occasionally be confusing.

It also has an author.

That is what separates much of the music I love. Thievery Corporation sound like Thievery Corporation. Mazzy Star sound like Mazzy Star. Massive Attack sound like Massive Attack. Everything but the Girl sound like Everything but the Girl. Tame Impala sound like Tame Impala.

“Alive Alone” does not sound like “Asleep from Day,” despite both involving The Chemical Brothers. The collaborators changed the direction of the music because the people involved actually mattered.

Modern commercial music can involve a performer, several producers, multiple songwriters, label executives, social-media teams, brand consultants and platform strategy. Every person may contribute something useful, but authorship becomes distributed across so many commercial interests that nobody is allowed to make a sufficiently strange decision.

The product becomes cleaner as the author disappears.

Familiar Music Feeds on Itself

The industry often claims that it is simply producing what people want.

That explanation ignores how preferences are shaped.

If somebody spends years hearing songs built around similar structures, production styles and emotional subjects, those patterns begin to sound normal. Anything outside them may initially sound awkward, slow or confusing.

The listener is then more likely to respond to familiar music because familiarity reduces the effort required to understand it. Platforms observe that response and promote more of the same. Labels see the platform performance and sign artists capable of fitting the established pattern.

The audience does not independently demand identical music. The industry, platforms and audience gradually train one another.

This is similar to recommendation algorithms in every other part of the internet. A platform shows somebody content similar to what they previously watched. The person watches it because it resembles something they already understand. The platform treats that behaviour as proof that the person wants an even narrower version of the same thing.

Eventually, preference becomes repetition.

This does not mean listeners are stupid or incapable of appreciating unusual music. Older songs repeatedly become popular again when younger audiences encounter them. “Fade Into You” has continued reaching people who were not alive when it was released.

That suggests audiences have not lost the ability to appreciate music that is slow, ambiguous or emotionally difficult. They may simply need to hear it outside a system constantly pushing whatever is easiest to recognise.

Generative AI Music Is Not Real Music

Generative AI music is where the factory reaches its most depressing conclusion. The industry spent decades treating breaths, mistakes, unusual voices, strange structures and artistic uncertainty as problems to be corrected, and generative AI has now arrived to suggest that the musician was another unnecessary problem.

I need to be honest about my own position here because I use generative AI. I have used it to create images, assist with writing, develop software and perform tasks that would otherwise require skills I do not personally have. It would therefore be hypocritical for me to claim that using generative AI automatically makes somebody lazy, talentless or morally bankrupt in every possible context.

I am not a photographer because I generated an image. I would not pretend that I personally arranged the lighting, positioned the camera, selected the lens and captured the subject. The image is a useful output that can support something else I created, such as an article, but the act of prompting it does not suddenly give me the abilities or experiences of a photographer.

The same distinction should apply to music. Somebody can generate a song, enjoy the result and use it as background audio, but they have not therefore become a singer, composer, instrumentalist or producer. They requested an output from a system that statistically reconstructed patterns found across music created by other people.

Generative AI music is not real music in the sense that matters to me because there is no human experience inside the performance. The voice did not belong to somebody attempting to control their breath. The singer did not become strained, nervous or emotionally overwhelmed during the recording. The instrumental did not emerge from musicians reacting to one another, and the lyrics were not written because somebody had lived through the event they were attempting to describe.

A human listener can still attach meaning to it. People can find emotional meaning in almost anything, including fictional characters, instrumental music, ambient noise and accidental sounds. The listener may hear a generated song during an important moment and remember it for the rest of their life.

That emotional reaction belongs to the listener. It does not prove that the machine placed genuine experience inside the recording.

This is where generated music differs from using technology to assist a musician. A singer using pitch correction still sang. A producer using software still made decisions about rhythm, structure and atmosphere. A musician using an artificial instrument still performed or composed something. Even when technology heavily transforms the result, there remains a person whose choices and limitations shaped the recording.

When somebody types a prompt and receives an entire song, most of that relationship disappears. They may choose between several outputs, alter the prompt or regenerate a section, but selection is not the same thing as performing, composing or producing the work from its underlying parts.

This does not mean every person must become an exceptional singer before they are allowed to express themselves through music. Some of the most memorable musicians are not technically perfect vocalists. Their limitations become part of the identity. Kevin Parker does not need to sound like a classically trained singer. Hope Sandoval does not need to overpower the instrumental. Beth Orton does not need every line to arrive with identical clarity.

The willingness to attempt it matters. A person can sing badly, record a strange beat, work with another musician, learn an instrument slowly or create something rough that only a few people will ever hear. That work may be imperfect, but the imperfection belongs to somebody.

There are singers, beatmakers and musicians everywhere trying to find people willing to collaborate with them. Generative AI makes it possible to avoid that entire human process because another person may be inconvenient, expensive, unpredictable or capable of disagreeing with the direction of the song.

That convenience is precisely why the result feels empty. Music has always involved friction. The voice does not behave perfectly. The instrument does not produce the exact sound inside your head. Another musician interprets the idea differently. A recording fails, and the attempt to repair it produces something better than the original plan.

Generative AI removes much of that friction, but the friction was not merely an obstacle. It was one of the places where identity developed.

The technology is also frequently promoted using quantity. A model can create hundreds or thousands of songs quickly, as though the world is suffering from a shortage of audio files. Streaming platforms already contain more music than any person could hear during their entire life. Producing millions of additional acceptable songs is not solving a meaningful cultural problem.

We do not need an unlimited quantity of songs that resemble music. We need individual songs that somebody had a reason to make.

That does not mean generative AI has no legitimate use anywhere in audio. It may help restore damaged recordings, remove background noise, isolate instruments, improve accessibility or help an existing musician experiment with sounds they then deliberately reshape. The boundary will not always be perfectly clean, especially as more creative software incorporates machine-learning tools.

The principle is still straightforward. Technology should assist the human author rather than replace the reason the work exists.

A generated song can imitate the sound of grief, love, addiction or desperation because it has encountered patterns associated with those emotions. It can assemble the correct words, chords and vocal inflections without ever experiencing anything resembling the emotion itself.

That may produce convincing audio. It does not produce a human performance.

The entire argument of this article is that great music creates a relationship between the experience of an artist and the experience of a listener. The artist does not explain everything. They leave fragments, breaths, images and emotional evidence. The listener then interprets those fragments through their own life.

Generative AI removes the experience on the other side. The listener can still fill in the blanks, but the blanks were not left there by somebody attempting to express something. They appeared because a model generated a statistically plausible imitation of human expression.

That is why I do not consider it real music. It can sound like music, be distributed like music and even become attached to somebody’s memories like music, but there is no musician inside it.

Humanity Was Never the Defect

The breath that arrives at the wrong moment is not necessarily a defect. The rasp in one line is not necessarily a defect. A lyric that takes years to understand is not necessarily a defect. A song that does not contain an obvious TikTok section is not necessarily a defect. An album that changes genre halfway through is not necessarily a defect.

Those qualities may be the exact reason somebody remembers the music.

A singer does not need to sound technically imperfect for the sake of sounding authentic. Deliberately making a recording worse would be another kind of performance. The point is not that every mistake must remain untouched.

The point is that the modern production system often cannot distinguish between an error that damages the song and an irregularity that gives the song an identity.

Anything unpredictable becomes something to correct. Anything difficult becomes something to simplify. Anything ambiguous becomes something to explain. Anything commercially risky becomes something to remove.

The result can sound flawless while feeling strangely empty.

There are modern musicians still resisting this process, and major artists can still release meaningful work. Tame Impala proves that somebody can use modern technology, reach an enormous audience and still construct a world that belongs unmistakably to them. Billie Eilish and other mainstream artists also demonstrate that recognisable ideas can survive inside the commercial system, even when the surrounding pressures remain visible.

The factory has not completely eliminated originality, and old music was never universally honest or brilliant. The problem is the direction of travel.

Music is increasingly optimised to be immediately understood, instantly distributed and rapidly replaced. The most memorable music often works in the opposite way. It reveals itself slowly. It changes as the listener changes. It allows uncertainty, contradiction and discomfort to remain inside the recording.

It gives the listener enough information to feel something but not enough to completely solve it.

That is why somebody can return to the same song after ten years and discover that it now means something else. The artist left blanks, and life gradually filled them in.

Modern music has more technology, more access and more talent than ever before. None of those things should make the music less human.

The danger appears when the industry begins treating humanity as noise.

The breaths, roughness, ambiguity, strange structures and unresolved emotions are not obstacles standing between the artist and a perfect recording. They are often the only evidence that an artist was ever inside it.