An Analysis of AI-Driven Deception, Royalty Dilution, and the Crisis of Authenticity in the Digital Music Economy

Executive Summary

The modern music industry stands at a critical juncture, defined by the confluence of three powerful and disruptive forces: the exponential rise of generative Artificial Intelligence (AI) in music creation, the structural vulnerabilities of the dominant “pro rata” streaming royalty model, and a fundamental shift in listener behavior towards passive, algorithmically-curated consumption. This report provides an exhaustive analysis of these interconnected trends, using the 2025 “Velvet Sundown” incident as a central case study to illustrate the systemic economic and cultural threats now facing human artists.

The investigation reveals that the digital music ecosystem’s core components—streaming platforms, media outlets, and audience perception—are critically unprepared for an environment where content, identity, and communication can be synthetically generated and manipulated at scale. The Velvet Sundown case, a multi-layered deception involving a fabricated AI band and a fraudulent spokesperson, successfully exploited these vulnerabilities, demonstrating how easily algorithmic virality can be achieved without provenance and how media verification can be bypassed.

Economically, the report deconstructs the “pro rata” or “market share” royalty system, arguing that its architecture inadvertently creates a perverse incentive structure. This model, which calculates per-stream payouts by dividing a platform’s total revenue by its total streams, is being systematically diluted by the massive influx of low-cost, high-volume AI-generated content. This creates an economic asymmetry that effectively subsidizes synthetic music at the direct expense of high-investment, human-created art. Projections from leading industry bodies like CISAC quantify this threat, forecasting that generative AI could cannibalize up to 24% of music creators’ revenue by 2028, representing a cumulative loss of €10 billion.

Psychologically, the report challenges the simplistic notion that listeners are merely apathetic to the origins of their music. A synthesis of academic research reveals a complex paradox: while listeners express a conscious bias against AI-labeled music, valuing “humanness” and “soul,” they often cannot distinguish AI from human compositions in blind tests and may even prefer the AI’s output. This preference for authenticity is fragile and easily circumvented by platform mechanics, such as mood-based playlists and a lack of content tagging, which de-emphasize artist identity and provenance. This shift is most pronounced in the rise of “functional music”—background soundscapes for focus, sleep, or relaxation—a category perfectly suited for anonymous, utilitarian AI generation that now competes in the same royalty pool as primary artistic works.

Finally, the report conducts a critical review of platform policies, particularly those of Spotify, finding a strategic distinction between combating illicit processes (bots) and permitting AI-generated content. This framework, combined with features like “Discovery Mode,” allows platforms to benefit from the engagement driven by a flood of synthetic media while publicly maintaining a stance against fraud.

The report concludes with a series of strategic recommendations for artists, rights holders, streaming platforms, and policymakers. It posits a future where the music market bifurcates into a high-volume, low-margin stream of functional AI content and a premium, high-value stream of human-centric art. The central challenge for the industry is to implement the economic, technological, and regulatory safeguards necessary to prevent the former from financially strangling the latter, thereby preserving the value of human creativity in the age of AI.

I. The Velvet Sundown Anomaly: A Case Study in Algorithmic Virality and Media Deception

The case of The Velvet Sundown in mid-2025 serves as a potent and alarming microcosm of the vulnerabilities inherent in the contemporary digital music ecosystem. It was not a single event but a multi-layered deception that successfully stress-tested and exploited weaknesses across the entire value chain, from platform algorithms and media verification processes to audience engagement patterns. By deconstructing this incident, it is possible to map the precise mechanisms through which synthetic media can achieve mainstream impact, manipulate public discourse, and challenge fundamental notions of authorship and authenticity.

1.1 The Emergence of a Ghost

In the late spring of 2025, a psych-rock “band” named The Velvet Sundown materialized on Spotify and began accumulating listeners at an astonishing rate. Within weeks, the act had garnered hundreds of thousands of monthly listeners, a figure that would soon approach one million. This rapid ascent was not driven by a traditional marketing campaign, a pre-existing fanbase, or critical acclaim. Instead, the band appeared to have emerged fully formed within the closed loop of Spotify’s recommendation engine, securing placements on several prominent and high-follower playlists.  

From the outset, the project was riddled with inconsistencies that signaled its artificial nature. The band’s press “photos” were comically unnatural, bearing the tell-tale signs of AI image generation, and their Spotify biography was a boilerplate, generic text devoid of any specific or personal information. Despite existing for only a few weeks, they were releasing music at a prolific rate, with a third album,  

Paper Sun Rebellion, announced just two months after their apparent inception. On social media, the band’s accounts engaged in a form of meta-commentary, pushing back against accusations of being AI-generated with cryptic statements like, “They said we’re not real. Maybe you aren’t either”. This combination of rapid, algorithm-driven growth and a complete lack of a verifiable real-world history established the core anomaly: a non-existent, synthetic entity was achieving platform metrics that would be the envy of many legitimate, hard-working human bands.  

1.2 The First Deception: The “Art Hoax” Narrative

As the band’s mysterious popularity grew, so did media curiosity. On July 2, 2025, Rolling Stone published a phone interview with a man identified as “Andrew Frelon,” a pseudonymous figure who claimed to be a spokesperson and “adjunct” member of The Velvet Sundown. In the interview, Frelon confirmed what many suspected: the band was not real. He framed the project as a deliberate “art hoax” and an act of “trolling,” admitting the music was created with Suno, a popular AI music generation platform.  

Frelon articulated a clear, if provocative, artistic and philosophical motive. “Personally, I’m interested in art hoaxes,” he stated, referencing historical examples of artistic deception. He positioned The Velvet Sundown as an exploration of the increasingly blurred lines between reality and fabrication in the digital age. “We live in a world now where things that are fake have sometimes even more impact than things that are real,” he explained. “And that’s messed up, but that’s the reality that we face now. So it’s like, ‘Should we ignore that reality… or should we dive into it and just let it be the emerging native language of the internet?'”.  

This narrative—a tech-savvy artist using AI to comment on the nature of authenticity—was compelling and fit neatly into the ongoing cultural conversation about AI. Frelon’s admission was widely reported, and for a brief period, the mystery of The Velvet Sundown appeared to be solved. The project was understood as a conceptual art piece designed to provoke debate, and Frelon was its creator and mouthpiece.  

1.3 The Second Deception: The Hoax Within the Hoax

The resolution was short-lived. Just one day after the Rolling Stone article appeared, the story took a sharp, bizarre turn. In a lengthy post on the platform Medium, Frelon revealed that his entire persona was a fabrication. He was not affiliated with The Velvet Sundown in any way; he had orchestrated a “hoax within a hoax” aimed squarely at the media.  

Frelon, whose real identity remains protected by his pseudonym, explained that he was a professional in web privacy who had experimented with generative AI in his own time. Observing the burgeoning story around The Velvet Sundown and noticing the lack of a coherent social media presence associated with the project, he saw an opportunity. “Suddenly, I had the crazy idea, what if I inserted an extra layer of weird into this story?” he wrote.  

He meticulously built a false identity, creating a fake Twitter/X account for the band that was more active and appeared more credible than the sparse, official-looking one linked in the band’s Spotify bio. He used this account to troll music writers and engage with journalists, eventually creating a fake Velvet Sundown Gmail account to handle interview requests. His digital footprint became so convincing that it superseded the actual, albeit minimal, presence of the true creators, leading to the interview with  

Rolling Stone. His stated motivation was not malicious but cautionary: “I write this with the intent not of shaming anyone… but in the hopes of inspiring a more careful approach to prevent the publication of blatantly false information by people with worse or more dangerous agendas than my own foolish experiment”. After the interview, the phone number he used was disconnected. This second layer of deception exposed a profound vulnerability not in the music platform, but in the media’s ability to verify sources in a purely digital, synthetic context.  

1.4 The “Real” Creators Emerge

The Frelon affair forced the actual creators of The Velvet Sundown to break their silence. A Twitter/X account, later confirmed as authentic by being linked in the band’s Spotify bio, publicly disavowed Frelon. “Someone is attempting to hijack the identity of The Velvet Sundown by releasing unauthorized interviews, publishing unrelated photos and creating fake profiles claiming to represent us,” a post from the account read. “We have no affiliation with this individual nor any evidence confirming their identity or existence”.  

In a subsequent email statement sent to media outlets, the creators further clarified their position. “Over the past days, a number of impersonator accounts have surfaced across social media platforms, publishing fabricated statements and AI-generated imagery falsely attributed to us,” the statement read. In a moment of deep irony, the creators of a project built on AI-generated music and synthetic identity took issue with being misrepresented by “fabricated accounts or synthetic media”.  

They defined their project not as a simple hoax, but as a “multidisciplinary artistic project blending music, analog aesthetics, and speculative storytelling”. While embracing “ambiguity as part of our narrative design,” they requested that reporting be based on verifiable sources. This chaotic, self-referential conclusion to the saga—where the originators of a synthetic creation demand authenticity from those reporting on it—perfectly encapsulates the disorienting nature of the current media landscape.  

Date

Event

Key Actors

System Vulnerability Exposed

Early-Mid 2025

The Velvet Sundown (TVS) appears on Spotify, rapidly gaining listeners.

The Velvet Sundown Creators

Platform Algorithm: Spotify’s recommendation and playlisting algorithms prioritize engagement metrics over artist provenance, allowing a non-existent band to go viral.

June 26, 2025

Music Ally reports on TVS’s mysterious rise.

Media, TVS Creators

Audience Perception: Listeners engage with and generate streams for a completely fabricated entity, driven by passive, playlist-based consumption.

July 2, 2025

Rolling Stone publishes an interview with “Andrew Frelon.”

Andrew Frelon, Media

Media Verification: A convincing but fake digital persona and social media presence can bypass the fact-checking processes of major journalistic outlets.

July 3, 2025

Frelon reveals his own hoax via a Medium post.

Andrew Frelon

Information Ecosystem: Demonstrates the ease with which a determined individual can inject false narratives into a developing story about a digital phenomenon.

July 3-4, 2025

The “real” TVS creators disavow Frelon and issue a statement.

The Velvet Sundown Creators

Cultural Paradox: Actors who leverage synthetic media for their own projects demand a standard of authenticity from others that they themselves subvert.

Export to Sheets

Table 1: The Velvet Sundown Incident Timeline

1.5 Algorithmic Complicity and Virality

The entire Velvet Sundown saga was predicated on its initial success, which appears to be a direct result of the structure of modern streaming platforms. When asked how the band achieved its staggering stream counts, Frelon offered a simple explanation: “I know we got on some playlists that just have like tons of followers, and it seems to have spiraled from there”. This points to the powerful feedback loop within Spotify’s algorithmic ecosystem. Once a track gains initial traction, whether organically or through other means, the algorithm may place it on larger playlists, exposing it to millions of passive listeners and causing its stream count to “spiral” upwards.  

This process is not necessarily fraudulent in the platform’s eyes. Glenn McDonald, a former data analyst at Spotify, commented on the incident by noting that the platform has “no protections” against AI bands and that from Spotify’s perspective, it is not even “clear that this is a bad thing to be ‘protected’ against”. The platform’s primary goal is user engagement; if an AI-generated track on a “chill instrumental beats” playlist keeps a user listening, it has fulfilled its function, regardless of its origin. This reveals a state of algorithmic complicity, where the system is not designed to differentiate between human and synthetic art, only between engaging and non-engaging content. The Velvet Sundown did not so much break the rules as it expertly played by them, exposing a system where virality can be detached from artistry, history, or even existence itself.  

II. The Pro Rata Paradox: How AI-Generated Volume Dilutes the Digital Royalty Pool

The controversy surrounding The Velvet Sundown is not merely a cultural or technological curiosity; it is deeply rooted in the fundamental economic architecture of the music streaming industry. The prevailing royalty payment system, known as the “pro rata” or “market share” model, was designed for a different era. In the age of generative AI, this model has transformed from a seemingly equitable distribution mechanism into a system with a structural flaw that inadvertently subsidizes low-cost, high-volume synthetic content at the direct expense of human artists. Understanding this “pro rata paradox” is essential to grasping the full scope of the economic threat posed by AI.

2.1 The “Market Share” Model Explained

The vast majority of major streaming services, including Spotify, Apple Music, and Amazon Music, utilize a market share payment system. The mechanics are straightforward in principle. Each month, the platform aggregates all of the revenue it has generated from two primary sources: paid subscriptions from premium users and income from advertisers who reach free-tier users. From this total revenue pool, the platform deducts its own share, which is typically around 30%.  

The remaining 70% constitutes the royalty pot that is to be distributed to all rights holders on the platform. This pot is then divided by the total number of streams across the entire platform for that month. This calculation yields a microscopic per-stream value. An artist’s or rights holder’s payout is then determined by their proportion of the total streams. For example, if an artist’s tracks account for 0.1% of all streams on the platform in a given month, they receive 0.1% of the total royalty pot. This model means that an artist’s earnings are not based on a fixed rate per stream, but on their competitive share of the total listening activity.  

2.2 The Economics of Dilution

The critical vulnerability of the pro rata model lies in its core mathematical relationship. The per-stream payout is not a constant; it is an inverse function of the total number of streams on the platform. The formula can be simplified as:Per-Stream Value=Total Platform StreamsTotal Royalty Pool​

An artist’s payout is then:Artist Payout=Per-Stream Value×Artist’s Streams

This structure means that if the total number of streams on the platform (the denominator) increases faster than the total revenue pool (the numerator), the value of each individual stream necessarily decreases for every single rights holder. This is the economic mechanism of dilution. When a small number of new, popular songs are added to the platform, the effect is negligible. However, when millions of new tracks are uploaded, as is now possible with generative AI, they can massively inflate the total stream count, thereby depressing the per-stream value for everyone, from global superstars to emerging independent artists.  

2.3 The Asymmetry of Production Cost

The pro rata model becomes particularly problematic when confronted with the radical asymmetry in production costs between human-made and AI-generated music. Creating a professional-quality song as a human artist is a resource-intensive endeavor, involving costs for studio time, session musicians, producers, mixing and mastering engineers, and marketing. This high-investment, high-effort process typically results in a relatively small number of tracks.  

In stark contrast, generative AI platforms like Suno and Boomy have demonstrated the ability to create millions of musical works at a marginal cost approaching zero. In May 2023, it was reported that Boomy alone had already generated over 14.5 million songs. These AI-generated tracks, even if they are of low quality or only attract a small number of streams each, collectively contribute to the massive inflation of the “Total Platform Streams” denominator in the royalty equation. A human artist’s single, high-investment track is therefore forced to compete for its share of the royalty pool against a tidal wave of no-cost synthetic content. This is not a level playing field; it is a system where the lowest-cost producer has a structural advantage in diluting the value of the entire market.  

2.4 Quantifying the Payout

The environment in which this dilution occurs is already one of extremely low margins for most artists. Per-stream payout rates vary significantly across platforms and are influenced by a complex web of factors. These include the listener’s geographical location (streams from wealthier regions typically pay more), their subscription status (a stream from a premium subscriber is worth significantly more than one from an ad-supported user), and the specific contractual agreements between the rights holder and the platform.  

The table below provides a snapshot of the estimated per-stream rates, illustrating the financial landscape.

Platform

Estimated Per-Stream Rate (USD)

Key Influencing Factors

Source(s)

Tidal

$0.0125 – $0.01284

High-fidelity focus, artist-friendly positioning, subscription tiers.

 

Apple Music

~$0.01

Primarily subscription-based, strong premium user base.

 

Deezer

~$0.0064

Mix of subscription and ad-supported tiers.

 

Spotify

$0.00318 – $0.00437

Large user base with significant ad-supported tier, pro rata model.

 

Amazon Music

~$0.00402

Integrated with Prime, multiple tiers.

 

YouTube Music

$0.0005 – $0.002

Heavily influenced by ad-supported vs. premium streams.

 

Table 2: Comparative Streaming Royalty Payouts & Influencing Factors

These figures demonstrate that even before the mass influx of AI content, artists required immense stream volumes to generate meaningful revenue. The pro rata model’s susceptibility to dilution means that this already challenging economic reality is set to worsen. The system, once a plausible method for distributing revenue based on popularity, has become a paradox: it now structurally favors the mass production of low-cost content, which in turn devalues the very pool of money it is meant to distribute, disproportionately harming the human creators it was ostensibly designed to support.

III. The Listener’s Dilemma: Authenticity, Apathy, and the Perception of AI-Generated Art

A central and often oversimplified argument in the debate over AI music is that listeners ultimately do not care about the origin of a song, only its quality or utility. The case of The Velvet Sundown, which garnered millions of streams from real human listeners, seems to support this view. However, a deeper investigation based on academic and psychological research reveals a far more nuanced and contradictory reality. The listener’s relationship with AI-generated music is not a simple binary of caring versus not caring. Instead, it is a complex psychological paradox where a conscious preference for human authenticity coexists with a subconscious enjoyment of synthetic output—a conflict that is largely mediated and exploited by the context of the listening environment itself.  

3.1 The Composer Bias

When listeners are explicitly aware that a piece of music is created by an AI, a clear negative bias emerges. Multiple studies have confirmed that people tend to report liking music less when they know it was composed by an AI. This “composer bias” appears to be rooted in a deep-seated skepticism about the capacity of a machine to create what is perceived as a fundamentally human art form. Listeners associate music with emotion, lived experience, and intentionality—qualities they do not believe an AI possesses. This leads to a reluctance to accept or enjoy AI-composed music, as it is perceived to lack the “soul” and authenticity of a human artist pouring their heart into a track. This has significant real-world implications, suggesting that if AI music were clearly labeled as such, it would face a considerable barrier to public acceptance.  

3.2 The Contradiction: Preference for Unlabeled AI

The most striking finding from recent research is how this conscious bias evaporates—and sometimes even reverses—when the origin of the music is obscured. In a key 2025 study by Lecamwasam and Chaudhuri, participants were exposed to both human- and AI-composed music under various labeling conditions (correctly labeled, incorrectly labeled, and unlabeled). The results were paradoxical. While participants consistently rated human-composed music as more  

effective at eliciting a target emotional state, they were simultaneously and significantly more likely to indicate a preference for the AI-generated music.  

This dissociation between perceived efficacy and stated preference is critical. It suggests that on a purely aesthetic or sonic level, the AI-generated tracks were often more immediately pleasing to the listener, even as the listener simultaneously believed the human tracks were doing a better job at conveying “real” emotion. Further supporting this, another study found that AI-generated music was able to fool approximately half of all listeners into believing it was human-composed. The implication is profound: the negative bias against AI is not an objective reaction to the music’s inherent qualities, but a subjective reaction to the  

label of “AI.” In a blind test, that bias disappears.

3.3 The Search for “Soul” and “Imperfection”

Qualitative data from these studies provides insight into what listeners are searching for when they try to distinguish human from machine. Participants consistently associate “humanness” with abstract qualities like “flow,” “realness,” “soul,” and, perhaps most interestingly, “imperfection”. The subtle variations, the slight deviations from perfect rhythm or pitch, and the organic ebb and flow of a performance are interpreted as markers of authentic human expression.  

Conversely, technical perfection can sometimes be perceived as a sign of a machine. Some listeners in the studies linked perfection with a lack of humanity, intuitively associating the flaws in a performance with emotional expressiveness. This search for an authentic, imperfect “soul” in music demonstrates that listeners do, in fact, care deeply about the human element. The dilemma is that their ability to reliably detect these markers is highly fallible. When AI-generated music was mislabeled as human, participants would often project these very same human qualities onto it, praising its “emotional nuance” or “uniqueness”.  

3.4 The Power of Framing

The decisive factor in how a listener perceives and values a piece of music is often the contextual frame provided to them. Research has shown that listeners’ aesthetic judgments are heavily influenced by information about the creator. For example, people report liking a piece of music more if they are told it was composed by Mozart compared to an unfamiliar composer, even if the piece is identical.  

The AI music studies confirm this “framing effect” with remarkable clarity. In the incorrectly labeled group of the Lecamwasam and Chaudhuri study, a significant number of participants preferred AI-generated music that they believed was human. This suggests that their preference was not for the music itself, but for the  

idea of human creation. The belief about the music’s origin is more powerful in shaping their appraisal than the actual sonic properties of the music.

The table below synthesizes the key findings from these crucial studies, highlighting the consistent pattern of context-dependent perception.

Study / Source

Key Finding on Preference

Key Finding on Perceived Efficacy

Impact of Labeling

Qualitative Themes Associated with “Humanness”

Lecamwasam et al. (2025)  

Participants were significantly more likely to indicate preference for AI-generated music.

Human-composed music was rated as more effective at eliciting target emotional states.

Preference is deeply tied to the perceived origin. Listeners preferred AI music they assumed was human.

“Soul,” “imperfection,” “flow,” “realness,” “organic,” “emotional nuance.”

Collins & Manji (2024)  

N/A

AI music fooled approx. 50% of listeners into believing it was human-generated.

Negative perceptions exist for music known to be AI, but this is moot if listeners can’t tell the difference.

Notions of “authenticity” and “emotional intentionality.”

Shank (2022)  

Listeners liked music less that they thought was composed by an AI.

N/A

The belief about the creator (e.g., AI vs. human, Mozart vs. unknown) heavily influences liking.

Skepticism about AI’s ability to have emotional responses or be central to identity.

Hong et al. (2022)  

Listeners appreciated AI music more if they perceived the AI generator as a “musician.”

N/A

Anthropomorphism (attributing human characteristics to the AI) increases acceptance and positive attitude.

“Authentic and emotional art forms.”

Table 3: Synthesis of Listener Perception Studies on AI Music

In conclusion, the assertion that “listeners don’t care” is a dangerous oversimplification. Listeners do care about authenticity, but their perception of it is malleable and highly susceptible to the context in which they encounter music. The core problem for human artists is not listener apathy, but rather that the dominant mode of music consumption—the decontextualized, algorithm-driven playlist—effectively “launders” the inauthenticity of AI-generated content, bypassing the listener’s innate preference for human creation and rendering it irrelevant.

IV. The Rise of the Functional Soundscape: AI’s Infiltration via Algorithmic Curation and Passive Listening

The economic dilution of the royalty pool and the psychological paradox of listener perception do not occur in a vacuum. They are enabled and accelerated by a profound cultural shift in how music is consumed. The traditional model of active, artist-centric fandom is increasingly being supplemented, and in many contexts replaced, by a mode of passive, utilitarian listening. This has given rise to the dominance of “functional music”—soundscapes designed for a purpose rather than for artistic engagement. This category, amplified by algorithmic curation, has created the perfect, legitimate, and massive pathway for AI-generated content to infiltrate the streaming ecosystem and compete for royalty shares.

4.1 From Active Artist Fandom to Passive Mood Management

Historically, music consumption was centered around the artist and the album. Listeners formed relationships with bands and creators, actively seeking out their new releases and engaging with their body of work as a cohesive artistic statement. While this model still exists, the streaming era has facilitated a monumental shift towards a focus on individual tracks curated not by artist, but by mood, activity, or context.  

This has led to the prevalence of what is known as “lean-back” or passive listening. For a significant portion of the day—during work, study, exercise, or commuting—many users do not engage with music as a primary activity. Instead, music serves as a background texture, a tool for mood regulation, or an aid for concentration. In these scenarios, the listener’s priority is not artistic discovery or emotional connection with a creator, but the fulfillment of a specific function. This behavioral shift creates a demand for music that is unobtrusive, consistent, and, crucially, anonymous.  

4.2 The Dominance of the Algorithmic Playlist

Streaming platforms like Spotify have not only catered to this shift but have become its primary architects. The most powerful tools for music discovery and consumption on these platforms are now algorithmic and editorially curated playlists. Personalized playlists such as “Discover Weekly” and “Release Radar,” along with activity-based playlists like “Deep Focus,” “Chill Vibes,” or “Beast Mode,” drive a colossal portion of all streams. Spotify has reported that its personalized playlists alone account for over 60% of total listening time on the platform.  

These playlists fundamentally alter the listener’s relationship with the music. They prioritize a consistent “vibe” or “mood” over artist identity. A listener turning on a “lo-fi beats to study to” playlist is not seeking to discover the next great hip-hop producer; they are seeking a specific sonic environment that aids concentration. The artists behind the tracks become interchangeable and invisible by design. This system primes users to accept music without provenance, as the playlist brand itself becomes the trusted curator, replacing the role once held by the artist, a radio DJ, or a music journalist.  

4.3 “Functional Music” as a Category

This ecosystem has given rise to a booming category known as “functional music.” Spotify itself defines this as content like “sleep sounds, light piano, meditation beats, and non-music noise”. It is music designed explicitly as a utility. This includes genres such as ambient soundscapes, white noise, nature sounds, binaural beats, and instrumental focus music. The value of this content is not in its lyrical storytelling, melodic complexity, or emotional depth, but in its ability to serve a purpose: to help someone focus, relax, sleep, or exercise.  

The demand for this content is immense, and it has become a significant segment of the streaming market. However, its proliferation has become a major point of contention within the industry. Some music distributors, such as Random Sounds, have now begun to refuse to distribute this type of content. Their reasoning is that platforms are already “full of this type of content, so much so that this already represents a big problem for the music industry”. The core issue they identify is that these generic, non-musical recordings are exploiting a royalty system built on copyright laws designed to protect musical compositions, not the sound of rain or a car horn.  

4.4 The Perfect Trojan Horse

Functional music represents the ideal Trojan horse for AI-generated content to enter and dominate the streaming market. It neatly solves the “listener’s dilemma” described in the previous section. In the context of functional music, the listener’s desire for human “soul” and “authenticity” is largely irrelevant. No one seeks the “lived experience” of the artist who recorded a waterfall or the “emotional vulnerability” behind a track of pure white noise. The content is judged solely on its utility.

This creates a market where anonymous, machine-generated content can compete on equal footing with, and often outperform, human-created content. AI is perfectly suited to generate infinite variations of these soundscapes at virtually no cost, tailored to the precise sonic parameters that algorithms have identified as optimal for focus or sleep. This content is not fraudulent; it is listened to by real people for a real purpose. Yet, it competes in the exact same pro rata royalty pool as a deeply personal song by a singer-songwriter or an intricately produced track by a band.  

The cultural shift towards passive, functional listening, amplified by the platform mechanics of algorithmic playlists, has therefore created a legitimate, massive, and non-fraudulent market that is structurally and philosophically biased towards anonymous, machine-generated content. This is not a niche or a loophole; it is a core feature of modern streaming behavior and the primary vehicle through which the economic dilution of the royalty pool is being executed on an industrial scale.

V. Economic Forecast: Quantifying the Cannibalization of Human-Generated Music Revenue

The systemic flaws analyzed in the preceding sections—the pro rata paradox, the manipulability of listener perception, and the rise of functional music—are not merely theoretical concerns. Their real-world financial consequences are now being quantified by major industry bodies, and the forecasts are stark. The data indicates that the music industry is not on the cusp of a problem, but is already in the midst of a significant and accelerating transfer of value away from human creators towards the producers of generative AI content. This process, termed “cannibalization,” threatens to fundamentally reshape the economic landscape for musicians.

5.1 The CISAC Projections

In late 2024, the International Confederation of Societies of Authors and Composers (CISAC), a global umbrella group representing over five million creators, published the first major global study on the economic impact of AI on the music and audiovisual sectors. The report’s conclusions are sobering. It projects that, under the current regulatory framework, generative AI could “cannibalize” or displace  

24% of music creators’ revenue by the year 2028.  

This is not a gradual erosion but a substantial and rapid decline. The study calculates that this will amount to a cumulative loss of €10 billion (approximately $10.5 billion) for music creators in the five-year period between 2023 and 2028. By 2028 alone, the annual losses are projected to reach €4 billion ($4.2 billion). These figures represent income that would have gone to human songwriters, composers, and performers in a world without the substitutional effect of mass-scale generative AI.  

5.2 Sector-Specific Cannibalization

The CISAC study provides a more granular breakdown of where these losses are expected to be most acute, and the data strongly corroborates the analysis of functional music and passive listening as the primary vectors for AI’s economic impact. The report predicts the following cannibalization rates by 2028 :  

• 30% of digital revenues (from streaming platforms) will be lost to AI-generated content.
• 22% of revenues from radio, TV, live, and background music will be lost.
• 21% of revenues from physical media like CDs and video will be lost.

Most tellingly, the report forecasts that AI-generated music will account for an incredible 60% of the music in B2B libraries by 2028. These libraries, which provide background music for commercial spaces like restaurants, cafes, and retail stores, are a prime example of a purely functional music market where businesses seek to reduce licensing costs and artist identity is irrelevant. This specific projection provides powerful quantitative evidence for the “functional music as Trojan Horse” thesis. Similarly, the study predicts that Gen AI music will account for approximately  

20% of traditional music streaming platforms’ revenues by the same year.  

5.3 The Value Transfer

The report is unequivocal that this is not a case of market disruption creating new wealth for all, but a direct transfer of economic value from creators to AI companies. While human creators face a cumulative loss of €10 billion, the market for generative AI music content is projected to experience exponential growth. The study estimates this market will explode from around €3 billion in 2023 to  

€64 billion in 2028. The revenues of Gen AI service providers in music are projected to rise from €0.1 billion in 2023 to €4 billion annually by 2028.  

This value transfer is driven by what the study calls the “unlicensed reproduction of creators’ works” used to train AI models, which then generate substitute products that erode the revenue streams of the very creators whose work provided the training data.  

5.4 Contradictory Projections and Nuances

It is important to acknowledge some market analyses that present a more optimistic, albeit potentially misleading, picture. Some reports project that the overall AI music market will grow to a staggering $38.71 billion by 2033 and that AI could boost total music industry revenue by 17.2% by 2025.  

This apparent contradiction can be reconciled by understanding the difference between the growth of the total market and the distribution of revenue within it. The total economic “pie” of the music industry may indeed get larger due to new applications, efficiencies, and markets created by AI. However, the critical issue highlighted by CISAC and other creator-focused studies is that the share of that pie allocated to human creators is projected to shrink dramatically. The growth will be captured primarily by technology companies and the producers of synthetic content. A 2024 study by Goldmedia reinforces this, warning that musicians could face a 27% revenue drop by 2028 if proper compensation systems for human creators are not put in place. The economic data, when viewed holistically, is not a standalone prediction; it is the quantifiable, lagging indicator of the systemic cultural and technological shifts that are already well underway.  

Metric

Projection Source

Projected Value / Impact by 2028

Key Details

Overall Music Creator Revenue Loss

CISAC

24%

The percentage of revenue at risk of being “cannibalized” by generative AI content compared to a scenario where AI doesn’t exist.

Cumulative Monetary Loss (2023-2028)

CISAC

€10 billion ($10.5 billion)

The total estimated loss of income for human music creators over a five-year period due to AI substitution.

Cannibalization of Digital Streaming Revenue

CISAC

30%

The projected portion of revenue from platforms like Spotify and Apple Music that will be displaced by AI-generated tracks.

Cannibalization of B2B Library Revenue

CISAC

60%

The projected share of the background/functional music market for businesses that will be composed of AI-generated music.

Growth of Gen AI Music Market

CISAC

€3 billion (2023) to €64 billion (2028)

The exponential growth of the market for AI-generated music and audiovisual content, showing a massive value transfer.

Projected Revenue Drop for Creators (Alternate)

Goldmedia

27%

A separate study warning of a similar scale of revenue loss for creators without the implementation of new compensation models.

Export to Sheets

Table 4: Projected Economic Impact of Generative AI on Music Creator Revenue (2023-2028)

VI. Platform Culpability and Policy Gaps: Navigating Spotify’s Stance on Synthetic Media

Streaming platforms, as the primary arenas where these economic and cultural battles are fought, are not neutral arbiters. Their policies, product features, and underlying business models actively shape the market, creating pathways and incentives that can either support or undermine human artistry. A critical examination of Spotify’s official stance and its platform mechanics reveals a carefully constructed framework that distinguishes between overt fraud and permissible synthetic content. This strategic ambiguity allows the platform to publicly combat manipulation while simultaneously benefiting from the massive influx of low-cost, engagement-driving AI content that dilutes the royalty pool for traditional artists.

6.1 The Hard Line on “Artificial Streaming”

Spotify’s public-facing policies are unequivocally strict when it comes to what it terms “artificial streaming.” This is defined narrowly as any stream that “doesn’t reflect genuine user listening intent,” specifically including automated processes like bots or scripts used to manipulate stream counts. The company emphasizes that it invests significant resources in detecting, mitigating, and removing this type of activity to prevent bad actors from illegitimately siphoning money from the royalty pool.  

The penalties for confirmed artificial streaming are severe. They include the withholding of all associated royalties, the correction of public stream counts and chart positions, and, in flagrant or repeated cases, the removal of content from the platform. As a further deterrent, Spotify has implemented a policy of charging labels and distributors a penalty fee per track when “flagrant artificial streaming is detected on their content”. This hard-line stance against fraudulent  

processes allows Spotify to project an image of itself as a protector of legitimate artists and a fair marketplace.

6.2 The Soft Stance on AI-Generated Content

This strict prohibition of bot activity stands in stark contrast to the platform’s far more permissive stance on AI-generated content. So long as the content does not violate other platform rules, it is generally allowed. In an interview, Spotify CEO Daniel Ek clarified that AI-powered tools for enhancing audio quality are permissible, but AI music that directly impersonates human artists is strictly prohibited. The primary guidelines for uploading AI-generated music are that the creator must hold the necessary copyrights and that the music must not infringe on third-party rights, such as by using copyrighted samples or lyrics without permission.  

This creates a critical and consequential loophole. An original song that is entirely generated by an AI platform like Suno, using no copyrighted training data and not impersonating a specific artist, is perfectly permissible under Spotify’s rules. If that song is then listened to by a real human—for example, as part of a “Deep Focus” playlist—that stream is considered legitimate. It is not “artificial streaming” by Spotify’s definition because it reflects “genuine user listening intent,” even if the music itself is entirely synthetic. This distinction between illicit process (bots) and permissible content (AI music) is the central pillar of the platform’s strategy. It allows a flood of AI-generated music to enter the ecosystem legitimately, where it then participates in and dilutes the pro rata royalty pool.

6.3 Discovery Mode: Monetizing the Algorithm

The platform’s role is not merely passive. Through features like “Discovery Mode,” Spotify has created a mechanism to directly monetize the algorithmic placement that drives so much passive listening. Discovery Mode allows artists and labels to identify priority tracks, and in exchange for accepting a lower royalty rate—effectively a “commission” paid to Spotify on streams generated within this context—the platform’s algorithm will increase the likelihood of those songs being recommended in personalized contexts like Radio and Autoplay.  

This feature turns algorithmic priority into a commodity. It creates a pay-to-play dynamic within the very playlists that are most conducive to passive consumption of anonymous music. An entity producing AI music at scale can use Discovery Mode to invest a portion of its potential earnings to boost the visibility of its tracks, accelerating the feedback loop of playlist placement and stream growth. This system further advantages high-volume producers who can play the numbers game, promoting a wide portfolio of tracks and amplifying those that gain algorithmic traction, regardless of their origin or artistic merit. It represents a direct financial incentive for the platform to promote content within these passive listening environments, further entrenching their dominance.

6.4 The Lack of Transparency and Tagging

Perhaps the most crucial element of the platform’s culpability is the deliberate lack of transparency for the end-user. Unlike other platforms like YouTube, which are moving towards mandatory disclosure, Spotify has no clear or enforced tagging system to identify synthetic or AI-generated content for listeners. The responsibility to disclose is left to the creator, a system that is inherently unreliable and unenforceable at scale.  

This absence of a label is not a neutral design choice; it is a decisive intervention. As established in Section III, the “composer bias” is only triggered when a listener knows music is AI-generated. By presenting all music in a uniform, decontextualized format within a playlist, Spotify prevents this bias from activating. It allows AI music to be consumed without the prejudice it would likely face if its origin were transparent. This policy of opacity is the final link in the chain, ensuring that the listener’s fragile preference for human authenticity is never given the chance to manifest, thereby allowing the economic dilution of the royalty pool to proceed unchecked. The platform’s architecture is not merely hosting content; it is actively curating a listening experience that makes the distinction between human and machine functionally invisible and, therefore, economically irrelevant.

VII. Strategic Recommendations and Future Outlook

The confluence of generative AI, the pro rata royalty system, and passive listening habits has created a systemic crisis that threatens the economic viability of human artistry in the digital age. The current trajectory, if left unaltered, points toward a future where the music market is flooded with low-cost synthetic content that devalues the work of human creators. Addressing this challenge requires a coordinated, multi-faceted response from all stakeholders. The following strategic recommendations are designed to mitigate the damage and realign the industry’s economic and cultural incentives to support human creativity.

7.1 For Artists and Rights Holders

Artists and their representatives must adopt a two-pronged strategy of defensive advocacy and offensive market positioning.

• Defensive Strategy: Advocate for Systemic Reform. The pro rata royalty model is the core economic vulnerability. Rights holders must collectively lobby streaming platforms to move away from a pure market-share system. Alternatives to advocate for include:
  • User-Centric Payment Systems (UCPS): Where a subscriber’s monthly fee is distributed only among the artists they actually listened to. This would sever the link that allows non-listened-to content to dilute the pool.
  • Artist-Centric Models: A hybrid approach that rewards artists for deeper engagement metrics beyond a simple 30-second stream. This could include higher weighting for track saves, playlist additions from fans, full album listens, and follows, thereby rewarding the creation of a dedicated fanbase over passive, fleeting plays.  
  • Legislatively, artists should continue to push for robust protections of their voice, likeness, and identity, following the model of Tennessee’s ELVIS Act, to prevent unauthorized AI cloning and deepfakes.  
• Offensive Strategy: Market “Humanity” as a Premium Attribute. In a market saturated with synthetic content, human authenticity becomes a scarce and valuable commodity.
  • Artists should lean into storytelling, direct fan connection, and live performance—areas where AI cannot compete. Building resilient communities around a verifiable human identity is the strongest defense against anonymous content farms.  
  • When AI is used, it should be as a transparently disclosed creative tool, not a replacement for artistry. This reframes AI from a threat to a collaborator, maintaining the artist’s control over the narrative and their work’s authenticity.  

7.2 For Streaming Platforms (e.g., Spotify)

Platforms must recognize that the long-term health of their ecosystem depends on a vibrant and sustainable community of human creators. Their current path risks creating a “slop” economy that may drive away both premium artists and discerning listeners.

• Mitigation Strategy: Implement Transparency and Segmentation.
  • Mandatory AI Tagging: Platforms must implement a clear, non-negotiable labeling system for all content that is primarily AI-generated. This would empower listeners to make informed choices and allow their natural “composer bias” to function, creating a market-based check on low-quality synthetic content.  
  • Separate Royalty Pools: To stop the direct cross-subsidy, platforms should consider segmenting content into different royalty pools. “Functional music” (e.g., white noise, rain sounds, AI-generated focus beats) should not compete for the same royalty dollars as primary artistic works. This would contain the dilutive effect of high-volume utilitarian content to its own category.
• Ethical Innovation: Realign Incentives with Artist Sustainability. Platforms should redesign their algorithms and royalty models to reward quality over quantity. An artist-centric model that prioritizes deep engagement, as described above, would realign the platform’s financial success with the success of its most valuable human creators, fostering a healthier and more sustainable artistic ecosystem.

7.3 For Policymakers and Industry Bodies

Government and collective organizations have a critical role to play in establishing a fair and lawful operating environment for AI in music.

• Regulatory Framework: Mandate Consent and Compensation. The most pressing issue is the use of copyrighted works to train generative AI models without permission or payment. Policymakers must establish clear international standards requiring that AI developers obtain licenses and provide fair remuneration to rights holders for the use of their work in training datasets. This would address the fundamental “theft dressed up as competition” that underpins much of the current AI economy.  
• Economic Intervention and Copyright Reform. The anticompetitive implications of the pro rata system in an AI-saturated market warrant investigation. Regulators should explore whether this model unfairly disadvantages smaller creators and facilitates the dominance of large-scale synthetic content producers. Furthermore, copyright law must be updated to provide clear guidelines on the ownership and rights associated with AI-assisted and AI-generated works, reducing the legal ambiguity that currently plagues the industry.  

7.4 Future Outlook

The music industry is irrevocably heading towards a bifurcated future. One stream will be a high-volume, low-margin, utilitarian world of functional soundscapes, algorithmically generated and passively consumed. This will be the domain of AI. The other stream will be a premium, high-value world of human-centric art, where storytelling, emotional connection, community, and verifiable authenticity are the core currencies. This will be the domain of human artists.

The existential challenge for the music industry is not to stop the first stream—it is likely too late for that—but to build the economic, technological, and legal firewalls necessary to prevent it from economically drowning the second. Without decisive action, the “pro rata paradox” will continue to drain value from human art to subsidize synthetic function, hollowing out the creative core of the industry. The strategies outlined above offer a pathway to a more balanced and sustainable future, one where technology serves as a tool for human creativity rather than its replacement, and where the profound value of human artistry is preserved and properly compensated.