Pith. sign in

REVIEW 3 major objections 5 minor 126 references

The Algorithmic Flattening of Sound: Computational Evidence and Justice Implications of AI Music Homogenization

T0 review · 3 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read An audit of 1,600 tracks shows Suno and Lyria 3 homogenize music in opposite but measurable ways.

desk verdict A careful audit of two commercial music generators with a genuinely new null-prompt design, but the central homogenization ratios rest on an instrumental-vs-vocal confound that the paper only partially controls. read the letter →

arxiv 2608.06106 v1 pith:HXLDBWTQ submitted 2026-08-06 cs.CY

Zoe Slendebroek , Danaé Metaxa This is my paper · ORCID
classification cs.CY
keywords AImusichomogenizationtext-to-musicgenerationinformationretrievalalgorithmicauditingculturaljusticegenreboundariesSunoLyria3
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper tries to establish that commercial text-to-music systems produce measurably homogeneous music, and that the homogenization is system-specific rather than a single 'AI sound.' Auditing Suno and Lyria 3 across Afrobeats, K-pop, Dance Pop, and Heavy Metal with 72 audio features, it finds Lyria shrinks within-genre acoustic variation (variance ratio 0.839) while Suno collapses acoustic distance between genres (separation ratio 0.442 vs. human 0.662) without narrowing within-genre spread. A standard classifier separates AI from human tracks almost perfectly (mean AUC 0.991) on these features alone, and the two systems are more distant from each other than random human subsamples (convergence ratios 2.70–4.64). The authors argue these patterns matter for cultural, economic, and epistemic justice because generated music increasingly flows through platforms that reward legible, predictable audio.

What carries the argument

The argument is carried by a 72-dimensional MIR feature space spanning rhythm and timing, spectral shape, MFCC timbral envelope, timbre and texture, harmonic content, structure and repetition, and dynamics, plus five complementary diagnostics: global dispersion (track-to-centroid distances), feature variance ratios, entropy ratios, PCA geometric coverage, and separability classification. The load-bearing quantities are three ratios—the aggregate variance ratio (AI/human within-genre variance), the genre separation ratio (between-genre centroid distance divided by within-genre spread), and the system convergence ratio (AI-system centroid distance divided by expected distance between random human splits)—together with the classifier's cross-validated AUC. These operationalize homogenization as reduced acoustic variation in a standardized space rather than as a subjective aesthetic judgment.

What would settle it

A re-audit using full-length human tracks (or a differently sampled human corpus) that found Lyria's variance ratio near or above 1.0 and Suno's genre separation ratio at or above the human 0.662 would directly contradict the claimed homogenization patterns; likewise, a classifier trained on matched instrumental human tracks across all four genres that dropped below roughly 0.9 AUC would weaken the claim that AI outputs are near-perfectly discriminable from human music on acoustic features alone.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that homogenization in AI music is not a single 'AI sound' but two structurally distinct tendencies that can be separated and measured. Using 72 music information retrieval features across four genres, the audit finds that Lyria 3 compresses the acoustic spread within each genre—aggregate AI/human variance ratio of 0.839, with 58% of feature-by-genre cells showing reduced variance—while Suno leaves within-genre spread intact or even enlarged (variance ratio 1.667) but collapses the acoustic distance between genres, lowering the genre separation ratio to 0.442 from the human 0.662. In addition, a random-forest classifier distinguishes AI from human tracks nearly perfectly on the same features (mean AUC 0.991 ± 0.003), with timbral dynamics and rhythmic regularity as the dominant cues, and this separability survives an instrumental-only Afrobeats control, indicating vocals are not the cause. The two systems are also more acoustically distant from each other than two random human subsamples (convergence ratios 2.70–4.64 across genres), so the findings point to learned, system-specific priors rather than a convergent 'AI default' or a prompt artifact.

Load-bearing premise

The human reference corpora—100 tracks per genre sampled from Spotify playlists via k-means and analyzed from 30-second Deezer previews—are representative of each genre's true acoustic range; if those playlists or previews omit genre-defining variation, every AI/human comparison is measured against a distorted baseline.

Editorial extensions

If this is right

  • If Lyria's within-genre compression generalizes, listeners streaming AI-heavy playlists of a genre will be exposed to a narrower acoustic band of that genre than human catalogs offer.
  • If Suno's genre-boundary collapse generalizes, the categorical identities that organize music libraries and recommendation systems—the separations that make Afrobeats and Heavy Metal distinct—become harder to maintain as AI tracks accumulate.
  • Because both systems' outputs are near-perfectly separable from human tracks on MIR features, acoustic fingerprinting of AI-generated music is feasible in principle, supporting disclosure, watermarking, or filtering—but also adversarial evasion if producers adapt.
  • The low prompt fidelity observed (Suno max |r|=0.26) implies that a user's attempt to steer generation toward a specific acoustic target will largely fail by default, so homogenization patterns reflect the systems' priors, not user choices.
  • If AI-generated music is especially legible to the classifiers and playlist algorithms that organize streaming platforms, generated tracks may be preferentially surfaced, creating a feedback loop that shifts genre reference distributions over time.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A testable extension would be a longitudinal audit: if human producers begin mimicking AI-typical timbral and rhythmic signatures to stay visible on platforms, the AI–human separation measured here should shrink over time, weakening detector-style watermarking that depends on a stable acoustic gap.
  • The near-perfect discriminability is measured on features that encode Western production norms; a listener study or a feature set built around microtiming and groove might find that the homogenization is partly an artifact of the measurement space, especially for Afrobeats and K-pop.
  • The authors' framework implies that the most consequential genre drift will occur in underrepresented genres, since those are learned from smaller and more Western-skewed samples; this could be tested by auditing additional non-Western genres such as Amapiano, reggaeton, or regional rap scenes.
  • If Deezer's 28% AI-upload figure is representative, the platform-feedback loop described here can be checked by measuring whether recommendation systems disproportionately surface AI tracks and whether that exposure shifts listener genre expectations in controlled experiments.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. This paper audits two commercial text-to-music systems (Suno v5.5 and Lyria 3) across four genres (Afrobeats, K-pop, Dance Pop, Heavy Metal), generating 100 tracks per system per genre under MIR-derived and minimal prompts, and compares their acoustic properties to 100 human tracks per genre using 72 MIR features and five homogenization diagnostics. The paper reports that Lyria compresses within-genre acoustic variance (aggregate AI/human variance ratio 0.839) while Suno expands it (1.667) but collapses genre-centroid separation (separation ratio 0.429–0.442 vs human 0.662), that the two systems do not converge (RQ3 ratios 2.70–4.64), and that a classifier separates AI from human tracks almost perfectly (AUC 0.991). It then develops a justice-centered interpretation of these findings in terms of recognition, redistribution, and epistemic justice.

Significance. If the empirical claims are supported, this is a valuable black-box audit: it compares two deployed systems against matched human baselines, introduces a minimal-prompt condition to separate system priors from prompt effects, includes an Afrobeats instrumental-only check for the classifier result, and clearly labels the normative conclusions as interpretive. The paper also makes good-faith efforts to audit the representativeness of the human corpora (geographic audit for Afrobeats, Billboard overlap for K-pop). The principal weaknesses are that the central homogenization ratios for RQ1/RQ2 are not tested under matched vocal/instrumental conditions and are reported without uncertainty quantification, which the revision should address.

major comments (3)
  1. [Methods (Experimental Designs, Human Reference Corpora); Findings (RQ1, RQ2); Limitations] The AI tracks in both experiments are generated with prompts that explicitly request instrumental, no-vocals audio (Methods, Experiment 1 and 2), whereas the human reference corpora contain vocals in three of four genres (Limitations states this explicitly). The paper's only matched instrumental control is the Afrobeats-only classifier run in RQ4; it does not recompute the within-genre variance ratios (RQ1) or genre-separation ratios (RQ2) on a matched instrumental human baseline. Since vocal content contributes substantially to acoustic variance and to genre-discriminative timbral features, the aggregate Lyria ratio (0.839) and the Suno separation ratio (0.429/0.442 vs 0.662) could partly reflect instrumentality rather than learned homogenization. The Afrobeats results (Lyria ratio 0.832) suggest compression survives matching in that genre, but the magnitude across genres and the Suno separation gap remain unquantified under matched conditions. Please rerun RQ1/RQ2 on instrumental human corpora (or otherwise equate vocal status) and report the resulting ratios.
  2. [Findings (RQ2, RQ3) and Appendix Table 5] The central claim that Suno collapses genre boundaries rests on a single separation ratio (0.429 in text vs 0.442 in Table 5; human 0.662) with no confidence interval, bootstrap, permutation test, or any inferential statistic. Likewise, the RQ3 convergence ratios (2.70–4.64) are reported as point estimates without uncertainty quantification. The paper reports permutation/bootstrap procedures for D1 in Table 3, but these are not applied to the RQ2/RQ3 ratios that carry the paper's structural claims. Please provide CIs and significance tests for the separation and convergence ratios, and reconcile the text/table discrepancies.
  3. [Methods, Audio standardization; Human Reference Corpora] The paper states "All analysis used 30-second excerpts from the middle of the track" for comparability, but the human reference corpora consist of 30-second Deezer previews (Human Reference Corpora). If the Deezer previews are not the middle 30 seconds, then all AI/human comparisons mix segment-location effects with system effects. Please clarify whether the human previews were aligned to the same segment location, and if not, quantify the sensitivity of the main ratios to preview position.
minor comments (5)
  1. [Methods, Experimental Designs] The "null-prompt" condition is not actually null because the prompts still contain "Instrumental ... No vocals, no lyrics"; consider renaming it "minimal prompt" or "genre-only prompt" to avoid confusion.
  2. [Findings, RQ1] The mixed-effects model that yields Cohen's d = −0.265 is not described; please specify the model formula, random effects structure, and how the variance ratio was computed (e.g., mean over all feature×genre cells).
  3. [Methods, Human Reference Corpora] The k-means cluster count k=10 is introduced without justification or sensitivity analysis; please report whether the main ratios are robust to reasonable choices of k.
  4. [Abstract and Findings, RQ1] The abstract's claim that Lyria reduces within-genre acoustic diversity is stronger than the data in Table 4, where Dance Pop shows a ratio of 1.119 (Lyria more diverse); please add a qualifier such as "in three of four genres" or "on aggregate."
  5. [General] No data or code availability statement is included; making the feature extraction pipeline and the anonymized feature matrices available would substantially strengthen reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the audit compares independently collected AI and human corpora, the classifier uses held-out folds, and the null-prompt experiment independently confirms the main pattern.

full rationale

The paper's central empirical claims are derived from direct comparisons between AI-generated tracks and human reference tracks that were collected independently from Spotify playlists and Deezer previews, not from the generative models' parameters or training objectives. The homogenization diagnostics (variance ratios, pairwise-distance ratios, and separation ratios) are descriptive statistics over these independent samples, so no result is an input to the measurement by construction. RQ4's classifier is evaluated with 5-fold stratified cross-validation and within-fold feature scaling, preventing the reported AUC from being a fitted-value artifact. Experiment 2 additionally uses genre-name-only prompts and reproduces the main homogenization pattern, which independently corroborates that the Experiment 1 MIR-steered prompts (derived from human tracks) are not what forces the findings. The paper's only author self-citation is Metaxa et al. (2021), used once to characterize black-box auditing methodology in the Limitations section; it supplies no numeric constant, no uniqueness claim, and no empirical result, so it is not load-bearing. The acknowledged vocal/instrumental mismatch between AI and human tracks is a potential validity threat that the paper explicitly discusses, but it is a confound, not a circular derivation: the measured ratios and classifier performance would remain well-defined quantities even if confounded. The normative and justice-centered discussion is labeled as interpretive and is not used to derive the measurements.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The audit introduces no fitted model parameters; the listed numbers are hand-chosen analysis choices in baseline construction and diagnostics. The principal assumptions are representativeness of human corpora and adequacy of MIR features. No new theoretical entities are introduced.

free parameters (3)
  • k-means cluster count for human corpus sampling = 10
    Hand-chosen to sample 100 tracks per genre proportionally from 400 candidates; changes which human tracks form the baseline.
  • entropy histogram bins (D3) = 20
    Hand-chosen bin count for Shannon entropy ratios; affects the redundancy diagnostic.
  • PCA components retained (D4) = up to 10
    Hand-chosen dimensionality cutoff for geometric coverage estimates; affects convex-hull volume ratios.
assumptions (4)
  • domain assumption MIR features adequately operationalize musical homogenization.
    The paper defines homogenization as reduced variation in 72 MIR features; if these features miss musically salient dimensions such as groove or microtiming, measured homogenization may not match culturally meaningful homogenization. Invoked in Methods: Feature and Homogenization Diagnostics; acknowledged in Limitations.
  • domain assumption Human reference corpora are representative of genre acoustics.
    All AI/human comparisons use 100 tracks per genre sampled from Spotify playlists and Deezer previews as the baseline. Playlist curation, artist caps, and 30-second previews could bias the baseline. Invoked in Methods: Human Reference Corpora; Limitations.
  • domain assumption Vocal/instrumental differences do not drive the main homogenization results.
    AI tracks are instrumental while human tracks may contain vocals in three genres; the paper controls the classifier with an instrumental Afrobeats subset, but does not re-run the RQ1/RQ2 homogenization metrics on matched instrumental-only corpora for all genres. See Limitations.
  • domain assumption Black-box outputs reflect system learned priors rather than prompt constraints.
    Low prompt fidelity and similar null-prompt results support the interpretation, but without access to model internals this causal attribution to learned priors is an interpretation. See Experiment 2 and Discussion.

how reviews work

0 comments
Cite this review

Pith. "Pith review of The Algorithmic Flattening of Sound: Computational Evidence and Justice Implications of AI Music Homogenization." pith.science (2026). https://pith.science/paper/HXLDBWTQ

@misc{pith2026260806106,
  author       = {Pith},
  title        = {Pith review of: The Algorithmic Flattening of Sound: Computational Evidence and Justice Implications of AI Music Homogenization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HXLDBWTQ}},
  note         = {Machine review of arXiv:2608.06106}
}
read the original abstract

This paper audits whether large-scale generative music systems exhibit measurable musical homogenization relative to human-produced music, and develops a justice-centered account of why this matters. We audit two commercially deployed systems (Suno and Lyria 3) across four genres (Afrobeats, K-pop, Dance Pop, and Heavy Metal). For each system and genre, we generate 100 tracks and compare them against human corpora of equal size, using 72 music information retrieval (MIR) features and multiple diagnostics of dispersion, redundancy, and separability. We define homogenization as reduced acoustic variation in standard computational audio features including rhythm and timing, timbre/spectral shape, and dynamics, both within genres and across genre boundaries. We also generate tracks using only a genre name as the prompt, with no additional instructions, to reveal each system's default musical tendencies. The results show two structurally distinct homogenizing tendencies. Lyria reduces within-genre acoustic diversity, while Suno collapses the acoustic distinctions between genres without compressing within-genre spread. Neither system follows user prompts faithfully, indicating that the observed patterns reflect learned priors rather than prompt constraints. The two systems do not converge on a common acoustic profile and are more acoustically distant from each other than two random human subsamples would typically be. Nevertheless, a standard classifier distinguishes AI from human tracks near-perfectly on MIR features alone. We argue that these patterns matter not as an aesthetic curiosity but as a justice-relevant condition, shaping which musical styles become legible, valued, and economically rewarded as generated outputs increasingly circulate at scale.

Figures

Figures reproduced from arXiv: 2608.06106 by the authors.

Figure 1
Figure 1. Prompt fidelity: Pearson r between encoded tar￾get feature value and AI output value (Experiment 1). Suno tracks prompts only weakly (|r| < 0.26 across all features and genres). Lyria shows moderate tempo tracking (r = 0.30–0.57) but limited fidelity elsewhere. RQ1: Suno Increases Within-Genre Dispersion; Lyria Homogenizes Within Genre The two systems exhibit qualitatively opposite patterns with respect to within-ge… view at source ↗
Figure 3
Figure 3. Genre separation ratio by system (RQ2). Higher values indicate more acoustically distinct genre categories. Suno’s separation ratio falls 36% below the human baseline, indicating that genre boundaries collapse in its output space. Lyria matches human genre separation (+1%). All ratios exceed 1.0, ranging from 2.70 (Heavy Metal) to 4.64 (Afrobeats), with Dance Pop at 4.38 and K-pop at 4.47. The two systems are theref… view at source ↗
Figure 4
Figure 4. System convergence ratio by genre (RQ3). Each bar shows the distance between the Suno and Lyria centroids divided by the expected distance between two random hu￾man splits. All ratios exceed 1.0 (dashed line), meaning the two AI systems are more acoustically different from each other than two random human subsets would be. RQ4: AI and Human Outputs Are Near-Perfectly Discriminable We train a classifier on the 72-dim… view at source ↗
Figures from the paper (2 more)
Figure 5
Figure 5. Figure 5: Top-12 features for AI vs. human discrim￾inability, Random Forest Gini importance (RQ4; AUC = 0.991±0.003). The three highest-importance features are MFCC ∆2 (timbral envelope acceleration, 0.144), MFCC 0 (overall energy envelope, 0.117), and IOI mean (inter-onset inte…
Figure 6
Figure 6. Figure 6: Word cloud of Suno null-prompt titles (100 tracks per genre). Word size proportional to frequency; color indicates the genre in which the word is most frequent (Afrobeats = amber, K-pop = pink, Dance Pop = blue, Heavy Metal = grey). The vocabulary is narrow and several…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

126 extracted references · 72 canonical work pages

  1. [1]

    Poetics , volume=

    And the hits just keep on coming: Music programming standardization in commercial radio , author=. Poetics , volume=. 2000 , publisher=

    2000

  2. [2]

    Studies in Philosophy and Social Science , volume=

    On Popular Music , author=. Studies in Philosophy and Social Science , volume=

  3. [3]

    2021 , publisher=

    The future of digital spaces and their role in democracy , author=. 2021 , publisher=

    2021

  4. [4]

    Art and Society , volume=

    The influence of african rhythms on modern music: A case study of afrobeat in Nigeria , author=. Art and Society , volume=

  5. [5]

    Proceedings of the 2023 AAAI/ACM Conference on AI, Ethics, and Society , pages=

    The ethical implications of generative audio models: A systematic literature review , author=. Proceedings of the 2023 AAAI/ACM Conference on AI, Ethics, and Society , pages=

    2023

  6. [6]

    Polity , volume=

    Race After Technology: Abolitionist Tools for the New Jim Code , author=. Polity , volume=

  7. [8]

    ILO Research Brief , number=

    Generative AI and the media and culture industry , author=. ILO Research Brief , number=

  8. [9]

    2024 , publisher=

    AI’s impact on music in 2025: Licensing, creativity, and industry survival , author=. 2024 , publisher=

    2025

Show all 126 references
  1. [10]

    K-Pop Artist 100 2025 , year =

    2025

  2. [11]

    Artificial intelligence, music recommendation, and the curation of culture , author=

  3. [12]

    Economic Inquiry , volume=

    Does digitization lead to the homogenization of cultural content? , author=. Economic Inquiry , volume=. 2022 , publisher=

    2022

  4. [13]

    The Guardian

    Music sector workers to lose nearly a quarter of income to AI in next four years, global study finds , author=. The Guardian. https://www. theguardian. com/music/2024/dec/04/artificial-intelligence-music-industry-impact-income-loss\#:\

    2024

  5. [14]

    2025 , publisher=

    Platform Jazz: Algorithmic Music Culture on TikTok , author=. 2025 , publisher=

    2025

  6. [15]

    2025 , howpublished =

    Music Glossary for. 2025 , howpublished =

    2025

  7. [16]

    2025 , howpublished =

    Music Generation Prompting Guide for. 2025 , howpublished =

    2025

  8. [17]

    Transactions of the International Society for Music Information Retrieval , volume=

    Data-Driven Analysis of Text-Conditioning in AI-Generated Music: A Case Study with Suno and Udio , author=. Transactions of the International Society for Music Information Retrieval , volume=

  9. [18]

    IASPM Journal , volume=

    Black on Black sounds: Music, Migration, and the ‘NU-K Blak’identity formation in early 21st century Britain , author=. IASPM Journal , volume=

  10. [19]

    IEEE Access , volume=

    Effective content recommendation in new media: Leveraging algorithmic approaches , author=. IEEE Access , volume=. 2024 , publisher=

    2024

  11. [21]

    2010 , publisher=

    Why voice matters: Culture and politics after neoliberalism , author=. 2010 , publisher=

    2010

  12. [22]

    Adjunct Proceedings of the Sixth Decennial Aarhus Conference: Computing X Crisis , pages=

    GenAI and the crisis of creative labor: Automation, augmentation, and the artist’s role , author=. Adjunct Proceedings of the Sixth Decennial Aarhus Conference: Computing X Crisis , pages=

  13. [23]

    Music Perception: An Interdisciplinary Journal , volume=

    There’s more to timing than time: Investigating musical microrhythm across disciplines and cultures , author=. Music Perception: An Interdisciplinary Journal , volume=. 2024 , publisher=

    2024

  14. [24]

    Marketing Science , volume=

    Changing their tune: How consumers’ adoption of online streaming affects music consumption and discovery , author=. Marketing Science , volume=. 2018 , publisher=

    2018

  15. [26]

    Future of Music Coalition, December , year=

    False Premises, False Promises , author=. Future of Music Coalition, December , year=

  16. [27]

    Social epistemology , volume=

    Conceptualizing epistemic oppression , author=. Social epistemology , volume=. 2014 , publisher=

    2014

  17. [28]

    2024 , howpublished =

    Dredge, Stuart , title =. 2024 , howpublished =

    2024

  18. [29]

    Creative Industries Journal , volume=

    Copyright, compensation, and commons in the music AI industry , author=. Creative Industries Journal , volume=. 2021 , publisher=

    2021

  19. [30]

    Information Sciences , volume=

    Rethinking data augmentation for adversarial robustness , author=. Information Sciences , volume=. 2024 , publisher=

    2024

  20. [31]

    , author=

    Attitudinal effects of mere exposure. , author=. Journal of personality and social psychology , volume=. 1968 , publisher=

    1968

  21. [32]

    2025 , eprint =

    Mehta, Atharva and others , title =. 2025 , eprint =

    2025

  22. [33]

    Policy Department for Justice , author=

    Generative AI and copyright: Training, creation, regulation. Policy Department for Justice , author=. Civil Liberties and Institutional Affairs, Directorate-General for Citizens’ Rights, Justice and Institutional Affairs, PE , volume=

  23. [34]

    Theory and Research in Education , volume=

    Bridging the gap between knowledge and justice: Epistemic challenges in participatory disability research , author=. Theory and Research in Education , volume=. 2025 , publisher=

    2025

  24. [35]

    New left review , volume=

    Rethinking recognition , author=. New left review , volume=. 2000 , publisher=

    2000

  25. [36]

    2025 , howpublished =

    Gomez Sarmiento, Ignacio , title =. 2025 , howpublished =

    2025

  26. [37]

    2024 , howpublished =

    Record Companies Bring Landmark Cases for Responsible. 2024 , howpublished =

    2024

  27. [38]

    , author=

    Exposure and affect: Overview and meta-analysis of research, 1968--1987. , author=. Psychological bulletin , volume=. 1989 , publisher=

    1968

  28. [39]

    Intellectual property office research paper forthcoming , year=

    Music creators’ earnings in the digital era , author=. Intellectual property office research paper forthcoming , year=

  29. [40]

    2022 , school=

    Creative mastering: A new culture of audio post-production , author=. 2022 , school=

    2022

  30. [41]

    Management Science , volume=

    Will the global village fracture into tribes? Recommender systems and their effects on consumer fragmentation , author=. Management Science , volume=. 2014 , publisher=

    2014

  31. [42]

    arXiv preprint arXiv:2010.05388 , year=

    AI song contest: Human-AI co-creation in songwriting , author=. arXiv preprint arXiv:2010.05388 , year=

    2010 arXiv

  32. [43]

    IFPI’s global study finds we’re listening to more music in more ways than ever , author=. Ifpi. Available online at: https://www. ifpi. org/ifpis-global-study-finds-were-listening-to-more-music-in-more-ways-than-ever/[Google Scholar] , year=

  33. [44]

    Royal Society open science , volume=

    Musical trends and predictability of success in contemporary songs in and out of the top charts , author=. Royal Society open science , volume=. 2018 , publisher=

    2018

  34. [45]

    Proceedings of the 16th Conference on Creativity & Cognition , pages=

    The impact of generative ai on artists , author=. Proceedings of the 16th Conference on Creativity & Cognition , pages=

  35. [46]

    Cultural anthropology , volume=

    Participatory discrepancies and the power of music , author=. Cultural anthropology , volume=. 1987 , publisher=

    1987

  36. [47]

    arXiv preprint arXiv:2510.20276 , year=

    From Generation to Attribution: Music AI Agent Architectures for the Post-Streaming Era , author=. arXiv preprint arXiv:2510.20276 , year=

  37. [48]

    Computer Law & Security Review , volume=

    Copyright protection during the training stage of generative AI: Industry-oriented US law, rights-oriented EU law, and fair remuneration rights for generative AI training under the UN's international governance regime for AI , author=. Computer Law & Security Review , volume=....

    2024

  38. [49]

    Journal of Visual Art Practice , volume=

    Rigorous creativity: AI, art and electronic life , author=. Journal of Visual Art Practice , volume=. 2025 , publisher=

    2025

  39. [50]

    2025 , howpublished =

    Marshall, Elizabeth Dilts , title =. 2025 , howpublished =

    2025

  40. [51]

    Royal Society open science , volume=

    The evolution of popular music: USA 1960--2010 , author=. Royal Society open science , volume=. 2015 , publisher=

    1960

  41. [52]

    Media, Culture & Society , volume=

    SoundScan and the consolidation of control in the popular music industry , author=. Media, Culture & Society , volume=. 1997 , publisher=

    1997

  42. [53]

    WIPO Magazine , year=

    How AI-generated songs are fueling the rise of streaming farms , author=. WIPO Magazine , year=

  43. [54]

    A Companion to Epistemology , volume=

    Testimonial Injustice and the Nature of Epistemic Injustice , author=. A Companion to Epistemology , volume=. 2025 , publisher=

    2025

  44. [55]

    Proceedings of the 2023 Conference on Human Information Interaction and Retrieval , pages=

    Why people skip music? On predicting music skips using deep reinforcement learning , author=. Proceedings of the 2023 Conference on Human Information Interaction and Retrieval , pages=

    2023

  45. [57]

    Foundations and Trends

    Auditing algorithms: Understanding algorithmic systems from the outside in , author=. Foundations and Trends. 2021 , publisher=

    2021

  46. [58]

    2025 , howpublished =

    Deezer Says 28\. 2025 , howpublished =

    2025

  47. [60]

    2013 , publisher=

    Music genres and corporate cultures , author=. 2013 , publisher=

    2013

  48. [61]

    2025 , eprint =

    Peeters, Geoffroy and Rafii, Zafar and Fuentes, Magdalena and Duan, Zhiyao and Benetos, Emmanouil and Nam, Juhan and Mitsufuji, Yuki , title =. 2025 , eprint =

    2025

  49. [62]

    American sociological review , pages=

    Cycles in symbol production: The case of popular music , author=. American sociological review , pages=. 1975 , publisher=

    1975

  50. [63]

    Social media+ society , volume=

    Locating power in platformization: Music streaming playlists and curatorial power , author=. Social media+ society , volume=. 2020 , publisher=

    2020

  51. [64]

    Fordham Intell

    No competition: How radio consolidation has diminished diversity and sacrificed localism , author=. Fordham Intell. Prop. Media & Ent. LJ , volume=. 2003 , publisher=

    2003

  52. [65]

    Rawls , author=

    A theory of justice. Rawls , author=. 1971 , publisher=

    1971

  53. [66]

    Scientific reports , volume=

    Measuring the evolution of contemporary western popular music , author=. Scientific reports , volume=. 2012 , publisher=

    2012

  54. [67]

    Arts , volume=

    Artificial intelligence and music: open questions of copyright law and engineering praxis , author=. Arts , volume=. 2019 , organization=

    2019

  55. [68]

    Audio Engineering Society Convention , volume=

    The loudness war: Background, speculation, and recommendations , author=. Audio Engineering Society Convention , volume=

  56. [69]

    International journal of music education , volume=

    The power of music: Its impact on the intellectual, social and personal development of children and young people , author=. International journal of music education , volume=. 2010 , publisher=

    2010

  57. [70]

    Emerging Economies Cases Journal , pages=

    Writers Guild of America Strike 2023 , author=. Emerging Economies Cases Journal , pages=. 2025 , publisher=

    2023

  58. [71]

    Stop Devaluing Music: An Open Letter to the Music Industry , year =

  59. [72]

    Copyright and Artificial Intelligence, Part 2: Copyrightability , year =

  60. [73]

    Adorno, T. W. 1941. On popular music. Studies in Philosophy and Social Science

    1941

  61. [74]

    Anderson, J.; and Rainie, L. 2024. The future of digital spaces and their role in democracy. Pew Research Center

    2024

  62. [75]

    Ayodele, T. 2024. The influence of African rhythms on modern music: A case study of Afrobeat in Nigeria. Art and Society, 3(1): 45--52

    2024

  63. [76]

    Benjamin, R. 2019. Race After Technology: Abolitionist Tools for the New Jim Code. Polity

    2019

  64. [77]

    Berardis, J. 2025. Towards responsible AI music: An investigation of trustworthy features for creative systems. arXiv:2503.18814

    2025 arXiv

  65. [78]

    Berger, V. 2024. AI's impact on music in 2025: Licensing, creativity, and industry survival. Forbes

    2024

  66. [79]

    Born, G.; et al. 2021. Artificial Intelligence, Music Recommendation, and the Curation of Culture. White paper

    2021

  67. [80]

    Bourreau, M.; Moreau, F.; and Wikstr \"o m, P. 2021. Does digitization lead to the homogenization of cultural content? Economic Inquiry, 60(1): 427--453

    2021

  68. [81]

    Burke, K. 2024. Music sector workers to lose nearly a quarter of income to AI in the next four years, global study finds. The Guardian

    2024

  69. [82]

    Burkhart, B. 2025. Platform Jazz: Algorithmic Music Culture on TikTok. Transcript Verlag

    2025

  70. [83]

    Suno. 2025. Music Glossary for Suno. https://help.suno.com/en/articles/9010177 Accessed: 2026-04-01

    2025

  71. [84]

    Google. 2025. Music Generation Prompting Guide for Lyria. Google Cloud Documentation. https://cloud.google.com/vertex-ai/generative-ai/docs/music/music-gen-prompt-guide Accessed: 2026-04-01

    2025

  72. [85]

    Casini, L.; et al. 2025. Data-driven analysis of text-conditioned AI-generated music: A case study with Suno and Udio. arXiv:2509.11824

    2025

  73. [86]

    Charles, M. 2025. Black on black sounds: Music, migration, and the ``NU-K Blak'' identity formation in early 21st century Britain. IASPM Journal, 15(2): 21--43

    2025

  74. [87]

    Chen, Y.; and Huang, J. 2024. Effective content recommendation in new media: Leveraging algorithmic approaches. IEEE Access, 12: 90561--90570

    2024

  75. [88]

    Choi, W.; and Koo, J. 2024. Large-scale training data attribution for music generative models via unlearning. arXiv:2506.18312

    2024

  76. [89]

    Couldry, N. 2010. Why Voice Matters: Culture and Politics After Neoliberalism. SAGE

    2010

  77. [90]

    R.; and Londre, T

    Danielsen, A.; Haugen, M. R.; and Londre, T. 2024. There's more to timing than time. Music Perception, 41(3): 176--198

    2024

  78. [91]

    Datta, G.; et al. 2017. Changing their tune: How consumers' adoption of online streaming affects music consumption and discovery

    2017

  79. [92]

    Dhariwal, P.; et al. 2020. Jukebox: A generative model for music. arXiv:2005.00341

    2020 arXiv

  80. [93]

    DiCola, P. 2002. A report on the effects of radio ownership consolidation following the 1996 Telecommunications Act. Future of Music Coalition

    2002

  81. [94]

    Dotson, K. 2014. Conceptualizing epistemic oppression. Social Epistemology, 28(2): 115--138

    2014

  82. [95]

    Dredge, S. 2024. Suno Releases Its First Mobile App after Attracting 12m Users. Music Ally, July 3, 2024. https://musically.com/2024/07/03/suno-releases-its-first-mobile-app-after-attracting-12m-users/ Accessed: 2026-01-13

    2024

  83. [96]

    Drott, E. 2021. Copyright, compensation, and commons in the music AI industry. Creative Industries Journal, 14(2): 190--207

    2021

  84. [97]

    Eghbal-zadeh, H.; et al. 2023. Rethinking data augmentation for adversarial robustness. Information Sciences, 654: 119838

    2023

  85. [98]

    European Parliament. 2025. Generative AI and copyright: Training, creation, regulation. Policy Department for Justice, Civil Liberties and Institutional Affairs

    2025

  86. [99]

    Felder, F. 2025. Bridging the gap between knowledge and justice: Epistemic challenges in participatory disability research. Theory and Research in Education, 23(3): 268--283

    2025

  87. [100]

    Fraser, N. 2000. Rethinking recognition. New Left Review, 3: 107--120

    2000

  88. [101]

    Gomez Sarmiento, I. 2025. AI-Generated Music Is Here to Stay. Will Streaming Services like Spotify Label It? NPR, August 8, 2025. https://www.npr.org/2025/08/08/nx-s1-5492314/ai-music-streaming-services-spotify Accessed: 2026-01-13

    2025

  89. [102]

    Hesmondhalgh, D. 2013. The Cultural Industries, 3rd ed. Sage

    2013

  90. [103]

    Hesmondhalgh, D.; et al. 2021. Music Creators' Earnings in the Digital Era. SSRN Electronic Journal

    2021

  91. [104]

    Hinksman, A. 2022. Creative Mastering: A New Culture of Audio Post-Production. PhD thesis, Birmingham City University

    2022

  92. [105]

    Hosanagar, K.; et al. 2014. Will the global village fracture into tribes? Recommender systems and their effects on consumer fragmentation. Management Science, 60(4): 805--823

    2014

  93. [106]

    IFPI. 2023. IFPI's global study finds we're listening to more music in more ways than ever

    2023

  94. [107]

    Interiano, M.; et al. 2018. Musical trends and predictability of success in contemporary songs in and out of the top charts. Royal Society Open Science, 5(5): 171274

    2018

  95. [108]

    Kawakami, R.; and Venkatagiri, S. 2024. The impact of generative AI on artists. In Proceedings of the Creativity and Cognition Conference, 79--82

    2024

  96. [109]

    Keil, C. 1987. Participatory discrepancies and the power of music. Cultural Anthropology, 2(3): 275--283

    1987

  97. [110]

    Kim, W.; et al. 2025. From generation to attribution: Music AI agent architectures for the post-streaming era. arXiv preprint

    2025

  98. [111]

    Li, K.; et al. 2024. Copyright protection during the training stage of generative AI. Computer Law & Security Review, 55: 106056

    2024

  99. [112]

    Manghani, S.; and Savage, T. 2025. Rigorous creativity: AI, art and electronic life. Journal of Visual Art Practice, 1--23

    2025

  100. [113]

    Marshall, E. D. 2025. Suno raises \ 250M funding led by Menlo Ventures, valued at \ 2.45B. Billboard

    2025

  101. [114]

    Mauch, M.; et al. 2015. The evolution of popular music: USA 1960--2010. arXiv:1502.05417

    2015 arXiv

  102. [115]

    McCourt, T.; and Rothenbuhler, E. 1997. SoundScan and the consolidation of control in the popular music industry. Media, Culture & Society, 19(2): 201--218

    1997

  103. [116]

    McEvoy, C. 2023. How AI-generated songs are fueling the rise of streaming farms. WIPO Magazine. https://www.wipo.int/en/web/wipo-magazine/articles/how-ai-generated-songs-are-fueling-the-rise-of-streaming-farms-74310

    2023

  104. [117]

    McWilliams, E. n.d. Testimonial injustice and the nature of epistemic injustice. PhilArchive. https://philarchive.org/archive/MCWTIA-2

  105. [118]

    Meggetto, F.; et al. 2023. Why people skip music? On predicting music skips using deep reinforcement learning. arXiv:2301.03881

    2023 arXiv

  106. [119]

    Mehta, A.; et al. 2024. Missing melodies: AI music generation and its ``nearly'' complete omission of the Global South. arXiv:2412.04100

    2024 arXiv

  107. [120]

    S.; Robertson, R

    Metaxa, D.; Park, J. S.; Robertson, R. E.; Karahalios, K.; Wilson, C.; Hancock, J.; and Sandvig, C. 2021. Auditing algorithms: Understanding algorithmic systems from the outside in. Foundations and Trends in Human--Computer Interaction, 14(4): 272--344

    2021

  108. [121]

    Musically. 2025. Deezer says 28\ Musically, September 11, 2025. https://musically.com/2025/09/11/deezer-says-28-of-its-new-music-uploads-are-now-ai-generated/ Accessed: 2026-01-13

    2025

  109. [122]

    arXiv:2508.11694

    Music and artificial intelligence: Artistic trends. arXiv:2508.11694

    arXiv

  110. [123]

    Negus, K. 2013. Music Genres and Corporate Cultures. Routledge

    2013

  111. [124]

    Peeters, G. 2024. Twenty-Five Years of MIR Research: Achievements, Practices, Evaluations, and Future Challenges. arXiv preprint. https://arxiv.org/html/2511.07205v1 Accessed: 2026-01-13

    2024

  112. [125]

    A.; and Berger, D

    Peterson, R. A.; and Berger, D. G. 1975. Cycles in symbol production: The case of popular music. American Sociological Review

    1975

  113. [126]

    Prey, R. 2020. Locating power in platformization: Music streaming playlists and curatorial power. Social Media + Society, 6(2): 1--11

    2020

  114. [127]

    Prindle, G. M. 2005. No competition: How radio consolidation has diminished diversity and sacrificed localism. Fordham Intellectual Property, Media & Entertainment Law Journal, 14(1)

    2005

  115. [128]

    Rawls, J. 1971. A Theory of Justice. Harvard University Press

    1971

  116. [129]

    Serr\`a, J.; Corral, A.; Bogu\ n\'a, M.; Haro, M.; and Arcos, J. L. 2012. Measuring the evolution of contemporary western popular music. Scientific Reports

    2012

  117. [130]

    Sturm, B. L. T.; et al. 2019. Artificial intelligence and music: Open questions of copyright law and engineering praxis. Arts, 8(3): 115

    2019

  118. [131]

    Vickers, E. 2010. The loudness war: Background, speculation, and recommendations. Available via Semantic Scholar

    2010

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.