There is a lot of talk about how a.i.-enabled tools will destroy music and the music industry – maybe even more conversations that the topic truly deserves. Most articles on “will a.i. replace real musicians” fail from the start by not defining what success means for musicians. Number of streams? Audience recognition? Getting an award? Artistic fulfillment?

Let’s state the obvious: generative a.i. tools certainly make it easy to generate the kind of homogeneous musical content that such tools were trained on. They are excellent tools for hobbyists to enjoy the process of generating an output that they (more and more comfortably) label as music. Their main features, convenience and ease of use, naturally come with a consequence: these tools excel at their ability to increase anyone’s productivity – as long as we measure “productivity” by quantity, speed and efficiency. And while the industry might use such metrics to measure success, it is not the primary criterion for most self-respecting artists.

Tech companies are the least impartial to express how musicians think and offer their advise on how creators can succeed. Some of their business narratives are so narrow-minded, demagogic or deeply misinformed (like this one), that reading them makes me feel frustrated and ashamed to be a human at the same time. I don’t have a direct stake in the a.i. game, but as a composer whose personal and professional life has revolved around the musical arts, and also has has been using, testing, developing, lecturing about music technology for over thirty years, I felt inclined to share my angle. The goal of this article is to examine everyone’s unique role in shaping the future of music and the music industry – even the role of the audience.

Trends: Music and the Music Industry

To begin, we must separate the concept of music and the music industry – which is highly relevant to separating how “success” in music is measured by different stakeholders. While music and music industry are often discussed together, these are quite distinct concepts; sometimes even seen as dynamically contrasting ideas. It is widely known among music professionals and even among today’s well-informed listeners, that behind the phrase ‘music industry”, more “industry” than “music” has been hiding for over a century. Most musicians argue that this uneven ratio has gotten even more out of balance over the past decades – and often blame the industry’s greed for it. (I don’t disagree about the detrimental effect of corporate greed, but would also add most creators’ egos to the core causes.)

The critics of this view regularly cite classic-era composers to prove their point, arguing that the “industry” aspect has always been part of the success of well-known composers. While it is true that networking, promotion, etc. were partial reasons behind their popularity, the music marketers of today often overlook some important details. First, Haydn, Mozart, Beethoven (or Bach, for that matter) might have been excellent service providers and promoters, but they also had to be quite good composers to begin with. It wasn’t their tools, rather their talent, skills, personal approach and experience that gave them the ability to create original, quality content. Most average composers, even if they had excelled at networking and promoting their average content, wouldn’t have survived in the music business of the 18th Century.

Secondly, the opposite is also true: there have been tons of original, high-quality content that didn’t see the light of day due to lack of promotion, keeping their creators hidden in the shadows of music history, yet, their lack of fan-base did not diminish the greatness of their work. Unlinking quality from popularity shows how the success of a composer measured in original and high quality output has little to do with the success of the same composer measured in financial goals and popularity. Therefore, the effects of a.i.-enabled tools on music vs the music industry should be also measured distinctly.

Change in Music Technology is Inevitable, but…

I happen to believe that it isn’t technology itself, but the ways new technologies are designed and encouraged to be used are what have resulted in the generation of more junk than even pop culture can swallow. Most (not all, but most) toolmakers focusing on profit via providing musicians with convenience and efficiency has been driving the direction of the music tech industry for at last 40 years. Whether we blame the designer or the user for this direction, a.i.-enabled tools are taking this to the extreme. They will erase some musician careers but might also be part of the solution for a much needed curation in the musical arts – a sort of natural selection. Let me explain.

In the 1960s it took as much (or more) skill to use a new guitar or a new synthesizer than a guitar or piano a hundred years before that. However, this can’t be said about the recent decades; too many tools since the 1980s have aimed at convenience, replacing talent and skill. For composers, performers, songwriters, such idiot-proof technologies have come in many forms (e.g. midi-loops, chord analyzers, auto-tune, trackspacer, auto-mixers, etc.). The democratization of distribution services (redesigned to support streaming, the most destructive model ever created), while beneficial for professionals, saturated the market with heaps of sub-standard content, diluting quality and the perceived value of music itself. (Spotify took this to the lowest level by enabling users to record any full-length album for 1.8 cents).

The most frequent argument I hear in favor or a.i. tools in music is that technology has always shaped music, whether it’s the evolution of electronic instruments or digital production tools. Some use cliche examples, like drum machines that didn’t replace drummers, sequencers that put keyboard players out of work, auto-tune that didn’t replace vocalists, and so on. These are quite poor comparisons that got old years ago. Where they all miss the point: virtually all of these past technological inventions changed the tools we used to create music, or effected how we used them. However, a.i. is not changing how a creator works, but replaces the ideators, the sources of musical concept, the creators themselves. It is not technology that reforms “the old ways” – or any ways; it is technology with the power to replace the user behind it.

Let that sink in for a second, please – this is why a.i. tools are in fact dissimilar to all disruptions we have dealt with before. Mix in the modern human’s desire for convenience, instant gratification and hunger for self-validation, and the result is a further shift of the (already upset) balance from carefully-crafted musical art to efficiently mass-made products.  But, there is also a silver lining.

Use or Resist a.i.-enabled Tools? It Depends on Your Goals.

Different people have different relationships with music. Their goals and the means to reach them can greatly vary as well.

I’ve discussed a.i.’s effect on music on various industry panels (most recently in AI vs the Audio and Music Professional in 2020 and in 2024, resp.) with diverse sets of experts. They included Maya Ackerman (computer scientist), Jonathan Wyner and Shahan Nercessian (iZotope), Alex Wankhammer (sonible) Nahuel Bronzini (mixing engineer), David Bowles (recording engineer), the late Steven Albini (record producer), Ken Felton (immersive audio technologist), Jean-Marc Jot (immersive audio technologist) and other professionals from music, technology or the music business. Back in 2020, we disagreed about the looming negative impact of a.i. on music, which became reality by 2024, the year of our second panel. Our second panel’s most pessimistic predictions became reality by mid-2025. Most tellingly, proponents of a.i. tools in music focused on benefits like convenience, business opportunities and the democratization of access (music as hobby), while panelists opposing the use of such technologies were more concerned with its negative effects on deeper human expression and ultimately, cultural standards.

I’m all for the democratization of tools of self-expression, but I do not support the false premise that new tools will turn casual users into artists. While I do not think that art should be reserved for a qualified few to create, I do believe that random products of casual self-entertainment should be distinguished from thoughtfully conceptualized & expressed works of art. Users who rely on cliche formulas to generate shallow imitations should not wear the same title (e.g. composer) as creators who invest in developing and refining their skills and commit to their craft for life. It’s time to acknowledge that creating music is not for everyone.

I was fully against a.i.-enabled music tech and saw no benefit to anyone back then. I understood the point of both its haters and supporters – but had three issues with it:

1. The suppression of originality / devaluation of the arts: generative music a.i. is built on the porridge made from the most promoted (therefore most over-played) music of the past, hence it serves up the cliches we had created over the past decades – or at best, repackages them as reusable building blocks. The inherent qualitative failure of a.i.-generated “music” lies in the way it’s constructed: the more widely such building blocks have been used (i.e. in more songs that became popular), the more likely they will end up in a.i.-generated content. It’s a self-feeding mechanism: it utilizes the most popular solutions, generates the safest, least innovative variations, then feeds the results back into the algorithm to fine-tune itself to offer increasingly more predictable popular variations. Do we really need more of the same aesthetically super-narrow cliches that labels have been showing down our throats for decades? I don’t think so – but the general audience who likes convenience and familiarity, might disagree. (This is also why the argument that a.i. algorithms learn audience preferences is a very backwards one.) In this aspect, the music industry wins, music and culture suffer.

2. Lack of concept, cultural context and personal expression: a.i. might be capable of generating entertaining content, even better quality content than some amateurish human-made music can be. However, a.i. does not “believe” in what it creates (unlike our best musicians). It might serve up logical solutions based on the context it is fed, but it has never experienced any cultural context or even artifacts. Where is the history, suffering, love, devotion, etc. – all that we experience – expressed in generated content? Personal interpretation and expression through art is what makes each artist unique – it’s not about logic and data, it’s about the expression of a concept formed by our individual lived experience. (And such personal expressions are even present in the millimeter-accurate and millisecond-timed decisions one makes when playing an instrument).

3. Emotional quantization: like in all digital systems, values are assigned on a stepped numerical scale in generative music. However small the steps are, there are always gaps between them (fractions of the smallest values in any code). The nuances of human emotions, and their expressions, are not based on values but are fluid and infinitely refined, “analog”. Just think of an already-nuanced feeling like bittersweet; each time you experience it, it is something different. Also, these multi-dimensional qualities are cross-dependent on one other (what type of bitter, what kind of sweet), dynamically influencing the outcome, often without logic. And then we haven’t even added the personal-contextual variables which further limit their assessment and imitation via code.

While I used to see these as “issues”, I see them now as reality that might actually lead to natural selection in today’s impostor-ridden music landscape. The question is: who will be effected and how.

Music and a.i.: Natural Selection

How about, for once, not ranking composers, musicians, songwriters, performers based on their position in the hierarchy of the music industry, but based on their level of artistic development? The ongoing changes are already reshaping the three fundamental creator segments:

• Hobbyists. The output quality of this suddenly expanded segment is getting close to the advanced creators’ output. Given that they see music as a delivery system for entertainment and/or for their self-expression, it is an opportunity for them and for the tech industry to offer them a.i.-supported services (e.g. writing assistance) – and with that, the illusion of possessing writing skills.

• Advanced creators. They are at the highest risk as the value of their advanced skills is diminishing. The a.i. music revolution is their new competition, so in order to keep their position, they are forced to adapt and utilize the new tools. Unfortunately, this comes at the price of (knowingly or unknowingly) lowering their standards as they their segment will devolve artistically. Otherwise their options are limited as they must cater to a pragmatic clientele or to an unassuming general audience. However, the most committed creators in this segment also have the opportunity to move “upwards” by developing into a top-tier artist (see diagram below).

•  Top-tier artists (defined here in terms of level of artistry, not revenue). For this narrow segment, a.i.’s effect is minimal; if anything, it’s pushing them to raise artistic standards. It is their suddenly accelerated opportunity to deepen their artistry NOW and evolve their individual style without a.i. (rather than re-invent themselves using a.i. – contrary to the generic advice of tech professionals). They only risk sliding down into the new abyss of well-packaged generic content if they get complacent with their own personal expression that has served them this far.

Maybe all this isn’t a bad thing for music overall. On the upside, the best will have to further improve, the mediocre will use a.i. to compete in a faster-paced, less-musical new segment. The respect of real music might return to where it was in the mid-20th Century, driven by an technological over-compensation for democratization. I think of it as a sort of music market correction. On the downside, the new music industry will likely flood the mainstream with an unlimited supply of a.i. mediocrity, regurgitated cliches and shallow imitations.

Natural selection and music market correction in an a.i.-infested music industry shift.

For conceptual originality and the expression of individuality, a human is required. The fate of the music industry is not in the creators’ and not even in the companies’ hands: it’s in the listeners’ hands. Whether that’s a good or bad thing, time will tell. Different listener segments expect creations at different artistic standards from different creators – therefore, every stakeholder has a different reason to adapt to the changing landscape.

Music is heading for an a.i. “Market Correction”

Since the 1980s, technology has made it “too easy” for many untrained writers working only on the surface of the quality and skill level that was previously expected in the musical arts. It has become the new norm to create seemingly original musical content that seemingly expressed deep layers. In reality, modern musical arts have become complacent with shallow imitations of the actual artistic depth and nuance found in sincerely designed and expressed works of top-tier creators. For them, music was never just a delivery mechanism but the core of their personality. Those who desperately relied on the assistance of technology due to their lack of skills, talent, artistic concept or commitment, quickly evolved into Advanced Creators (see the segment above). Conversely, the same technology only enhanced real artists’ concepts and sincere expressions, representing the best of the modern music culture.

However, over time, the lines got blurred between these segments; as advanced creators claimed their place in the top-tier artist segment, the general audience was getting heavily conditioned by packaging (production) and popularity (industry). The result: the majority of today’s listeners think that the more products an artist sells, the better they are at their craft (i.e. top-tier creators are likely with major labels). Yet, reality shows us that there is close to no correlation between a musician’s artistic quality and their position in the hierarchy of the music industry.

Now with a.i. bringing generic solutions and leveling the playing field for all but the truly original outstanding individual creator, top tier artists must work harder to separate themselves from the new, unified segment of mediocrity. While the advanced creators understandably fear a.i.-tools’ impact on their position, risking to be consumed by the new segment of mediocrity, they are faced with the true value of their art, their own humility and commitment to improve it.

How? Sincere and committed creators can grow by seriously improving the originality of their concepts, the individuality of their expressions, and the contexts within which they create – the areas a.i. tools are at an inherent disadvantage.

However, another important step for the new top-tier of artists to prevail is the elevation of the audience’s standards. This is not just an opportunity for them but for the the musical arts and culture. This can only take place through education – which is possibly the biggest challenge, given our recent trends of sinking appreciation of the arts and pragmatic priorities. The a new audience, the task is not unlike learning what to appreciate in a Monet painting beyond the obvious. Characteristic brush strokes are an easily observable, learnable and even imitable technique – relatively easy tasks for the general public or for an advanced a.i. engine. But, internalizing the painter’s personal intent with making the painting and every bit of passionate detail and emotional context behind each brush stroke on it takes far more attention, context, and a deep sensitivity – certainly more than code ever will be capable of. It requires a first-hand experience of the human condition.