Art Official Intelligence: Music and Intellectual Property Philosophy in the New Age of Artificial Intelligence–Assisted Creativity

3.1 Incentives‑based utilitarian argument

The incentives‑based argument for IP controls is built on the principles of utilitarianism, generally stemming from the formulation developed by J. S. Mill (Mill, 1863). In its most well‑defended modern line of reasoning, utilitarianism advocates courses of action that lead to the greatest positive impact on human flourishing. This suggests an initial presumption against protections on intellectual creations, because these would serve as barriers to understanding and to new creation; locking up culture hinders education and access to ideas, which are essential to utility and progress. However, insufficient incentive could reduce the amount or quality of creative generation in a society, which would impede utility or flourishing, and so some sort of legal mechanism to encourage or reward useful creativity can be justified (Moore, 2008). Economic analyses of IP’s effects (Landes and Posner, 1989) typically proceed from an implicitly incentives‑based utilitarian standpoint. IP systems based in utilitarianism would, in principle, govern with a light touch. They would offer limited controls over the works and for limited time periods, protecting only enough to encourage creation but not so much as to hamper the dissemination of knowledge and culture for the improvement of society in general (Vaidhyanathan, 2001). This rationale undergirds much of the Anglo‑American IP tradition (Goold and Simon, 2024); for instance, the US Constitution explicitly states that copyright and patent controls are to be implemented ‘to promote the Progress of Science and useful Arts.’ However, the large‑scale expansion of copyright controls worldwide since the 1970s has tended to cut against this bedrock principle, and some theorists have used versions of the incentives‑based argument to argue that many legal regimes now overprotect creative works to the detriment of society (Lessig, 2004).

3.2 Lockean property‑based argument

The Lockean theory of IP argues for a natural right in one’s creations by extending John Locke’s theory of physical property (Locke, 1689) into the realm of ideas and their expressions. Locke’s original conception began with the assertion that one owns one’s own body and, therefore, the work that the body does. When work is exerted on unowned objects and something new of value is fashioned, Locke held that the created value rightfully belongs to the laborer, and that state protection of that right is also thereby justifiable. Modern Lockean theorists have argued that intellectual work—creative thought (Moore, 2012) or time and mental energy spent developing new intellectual objects (Himma, 2012)—engenders analogous rights and thus merits similar protections. Historically, the Lockean justification has not had a strong practical impact on legal regimes, though it is typically introduced as a first principle in IP scholarly studies as well as for property theory generally (Merges, 2011). Lockean theory does not inherently suggest particular practical controls; it can be used to advocate for strong controls, as an outgrowth of more general property rights of use and exclusion (Mossoff, 2003), or for relatively limited ones, to allow room for competing claims based in liberty or the avoidance of waste and spoilage (Gordon, 1993). Nevertheless, as Boyle (2008) notes, modern Lockean thinking has tended toward expansive conceptions of the role of IP law.

3.3 Personality‑based argument

The third common foundation of IP is also based in a theory of natural rights, but, rather than focusing on labor, it foregrounds the concepts of personal autonomy and expressive individuality. The modern personality‑based argument begins with ideas about creativity based in French and German enlightenment philosophy, especially the work of Immanuel Kant and G.W.F. Hegel, though the popularization of the idea in the 19th century was accomplished through outspoken defenders in the realms of literature and commentary, such as Victor Hugo and Mark Twain (Kretschmer and Kawohl, 2004). Personality theorists hold that intellectual protections are necessary to allow for the development of each person’s unique creative voice. On this view, authors have inherent rights to communicate their own personalities, emotions, and experiences, which they infuse into the creative works they bring forth. This allows creators to become fully autonomous beings, able to speak for themselves on their own terms and governed by their own will (Joffrain, 2001). This argument is the intellectual basis for the protections known as moral rights, which can include the right to withdraw a work from circulation or the right to object to alteration of a work (Joffrain, 2001; Rajan, 2011). Many continental European IP regimes are justified through personality‑based principles; correspondingly, they have tended historically to afford longer and stronger protections and have afforded creators moral rights that are generally absent in nations that base their IP laws on Anglo‑American models. Rights of publicity, which are external to copyright but which protect individual creators’ personalities vis‑a‑vis commercial uses of artistic works, also tend to be stronger in countries with personality‑based traditions such as France and Germany than in utilitarian‑based ones such as the United States, where there is no federal right of publicity (Juzon, 2024). The theory has faced criticism from observers who note that it assumes a model of creativity that is a poor fit for some types of media (Hughes, 1988), such as most nonfiction, trivial Internet postings, and many works of corporate or collaborative authorship.

4.1 Incentives‑based utilitarian

At first glance, AI tools may appear to be a blessing to utilitarian theorists. These new capabilities unlock new avenues for creation; they open up creative activities to people who did not have the ability or skill beforehand; they help unearth new knowledge and reveal creative connections as yet undiscovered by humans. Concerns over how algorithms might distance creators from their creations are largely irrelevant to the utilitarian; the value of the creations’ effects is the primary benchmark (Kaminski, 2017). Whether and how AI will contribute to human flourishing generally—a deeply uncertain question at current—is beyond the scope of our paper; nevertheless, the capacity for the generation of potentially copyrightable works, in music as in other areas, is vastly increased by AI’s rise, suggesting a future in which there is an almost limitless supply of creative output (Desai and Lemley, 2022).

The crucial factors for utilitarians will be AI’s effect on incentives and the positive or negative impact of the works created by means of it. An increase in the quantity of creative works by no means equates to an increase in the quality of creative effort. Artificial creation may flood the marketplace with uninspired, ‘good‑enough’ musical sounds that drive human creators away from innovation, negatively impacting both the ability of composers and performers to make a living and general cultural development. Indeed, it seems almost inevitable that AI will come to supplant musicians for basic functional music needs, like filler music for corporate videos, advertising soundtracks, background music in public spaces, and music for exercising or meditating; these are milieus where composers and performers are likely to be pushed out because the labor cost far outstrips the cost to automatically create generically suitable music. This may result in a serious drain on incentives that threatens entire wings of socially beneficial creative activity.

It is worth noting, however, that musicians and the music industry have been forced to adapt to new technosocial paradigms routinely since the introduction of recording technology in the late 19th century. The commercialization of the phonograph, radio, the long‑playing record and compact disc, music television, file‑sharing, and music streaming have all disrupted and restructured musical creation, but they have also suggested new paths forward for creativity and economic gain (Singer and Rosenblatt, 2023). The music industry as a whole may need to retool to adapt to AI’s coming prominence, and individual musicians may find their work processes changing, but it is far from certain that the ability to receive incentives for music‑making will crumble under the weight of AI. Utilitarian theorists might therefore focus on sculpting the techno‑legal landscape in order to preserve and improve the quality of created content, such as by ensuring factual accuracy and supporting uses that foster greater expressive force.

4.2 Lockean

The Lockean labor‑property theorist may start by noting that there is not much labor in the work of AI; it is, in many cases, a labor‑saving tool, which suggests a prima‑facie argument that it lessens the bond between creator and work and, therefore, undermines the claim to a natural right of property in the artistic objects created. Simply by suggesting a vague idea or topic and a musical style, a user can now have an AI algorithm like Suno or Google’s MusicLM generate entire compositions. How can one argue a natural right in the fruits of labor that requires so little skill, so little intellectual input and expertise?

That said, some AI creativity may require sophistication and pre‑existing mastery of concepts in order to manipulate the AI tools into producing something aesthetically valuable. The practice of ‘prompt engineering,’ for instance, has strong Lockean components to it—it takes intellectual work to fashion prompts so as to get specific artistic results from AI tools (Chang et al., 2023). This reasoning has recently been rejected in the United States (Register of Copyrights, 2025a) because prompt‑writing is a sequence of (uncopyrightable) input ideas, whereas the (potentially copyrightable) expressions—the resulting AI artistry—is not fundamentally determined by the prompter. This points to a second problem for Lockean theorists: the black‑box nature of many AI tools. The processes by which these tools operate are often not known, and the output is not necessarily traceable to its inputs; in fact, outputs can differ dramatically even with little to no change in the inputs (Register of Copyrights, 2025a). The Lockean might therefore remark that the laborer’s investment of intellectual work is not necessarily resulting in a deliberate, directed creation that reflects the laborer’s own intentions. Mere coincidence makes for a weak property claim. Perhaps, however, such creations are akin to gold found on a unified‑estate cattle ranch; the farmer’s possession of the land legitimates the claim to the gold, even though mineral wealth was not the rationale for the establishment of the land title. On this metaphor, the AI‑aided creator could still make a defensible assertion of a property right. Work done by the human after the AI has finished, moreover, may provide stronger claims on behalf of prompters to true Lockean authorship.

4.3 Personality‑based

The growth of AI creation is seemingly at the strongest of odds with the personality‑based philosophy. This is art with no art, with no soul, no true human investment or creative spark; it is the exact opposite of what the personality‑based theory tries to safeguard, a nightmarish literalization of the ‘robot pop’ aesthetic adopted by electronic pop pioneers Kraftwerk. From the personality‑based perspective, AI‑shepherded creation seems to stand as a direct threat to the Romantic idea of creative genius or to auteur theory in film, both of which have strong roots in the personality‑based intellectual lineage (Kaminski, 2017; Rajan, 2011). The ability to create unauthorized digital replicas like ‘Heart on My Sleeve’ is particularly noxious to a personality theorist, and even allowing the works of unwilling or unwitting human creators to be used in AI datasets may be objectionable to artists who are committed to the personality‑based position. AI approximation of artists’ general patterns of expression further muddies the already‑murky waters of ‘substantial similarity,’ the legal test for determining when the replication of a creative style bleeds into infringement of specific works (Sobel, 2024). The sheer volume and velocity of AI‑fueled creation (Lemley, 2024) indicates that creators may find it fundamentally difficult to legally defend their unique expressions against imitation products in this new environment.

One possible saving grace, at least for some works, is that AI assistance may be necessary for certain creators to more fully implement personal autonomy. Creators with disabilities, or grand thinkers who seek to make large‑scale multimedia works that stretch beyond the bounds of their narrow expertise, could find in AI a way to fulfill their aesthetic visions. Personality theorists may also advocate for a circumscribing of AI’s reach; perhaps the mass of run‑of‑the‑mill works are suitable for machine‑based creation and transformation, while a subset of works reflecting deep human experience would be walled off from the AI world and given additional protections commensurate with a belief in the works’ inimitable expressive spirit. Carve‑outs in allowances for AI text and data mining have already been proposed, both for economic and personality‑based reasons, in countries such as Canada (Craig, 2025) and the United Kingdom (Kyle, 2025).

5.1 Incentives‑based utilitarian design

Most current AI tools take implicitly incentives‑based approaches in their design, working within the paradigms of US copyright law with the intention of maximizing creative fruition. Calls for AI design that center around making it easier for people to create (i.e., ‘democratizing’ creation) (Tang et al., 2024), as well as lowering barriers for making use of past works to inspire new creations, align with incentives‑based utilitarian thinking. At the same time, there is apprehension that AI tools may demoralize human creators, especially if AI creation overwhelms the market with low‑quality works (Lovato et al., 2025); an AI‑flooded market could disincentivize people from creating and sharing truly new and innovative works. Design built within the principles of incentives‑based utilitarianism will aim to ensure that the tools encourage users to generate new works and disseminate them widely but also strive to adequately reward creators so as not to deter the creative drive. Tools championing this approach should be built for sharing, collaboration, and compensation.

Build for Sharing. An important aspect of the incentives approach includes the ability for one’s work to be used and consumed by others. In order to do this, systems should be built such that they not only allow but encourage the sharing of AI creations—both during and after the creative process. This complements a shift from the view of art as a solitary act to one embedded in networks of creative use (Chung et al., 2022). Designers could support specific modes of sharing by modeling the types of informal information‑sharing practices that communities employ and integrating them into the interaction capabilities of the tools (Chung et al., 2022; Shneiderman, 2008). This can help creative ideas proliferate from the ground up and hasten the process of mutual inspiration. Current AI tools often already have sharing features; Midjourney’s Explore page and Suno’s library page offer possible models for dissemination.

Build for Collaborative Creativity. Tools that support cross‑collaboration among different AI functions and multiple human creators could spur inventive new creative uses. When artists can use technology to influence and expand each other’s efforts beyond their individual abilities, new and exciting creative outcomes may result (Amabile, 2018). AI tools can be built to promote serendipitous creations and iteration (Andersen and Knees, 2016; Huang et al., 2020), especially creations beyond a musician’s familiar creative territory (e.g., composition in disparate musical styles or with unusual instrumentation).

Literature related to collaborative creativity and music creation frequently notes the desire to be able to iterate through ideas. This may come in the form of conversation‑based interaction (Choi et al., 2025) allowing users to create a dialogue with AI systems, AI chatbots playing a mediating role in interactions between human collaborators, or in the construction of AI systems capable of simulating personas and roles taken within the creative process (Fu et al., 2025).

Taking into account communities of practice during the design process can also serve to support collaboration. AI tools that are crafted to accommodate more niche creative networks can encourage the use and sharing not only of the content of AI but also of creative processes and reasoning as well. The integration of AI tools into already existing workflows in communities (Morris et al., 2025) can help those communities reach their artistic goals, which tracks broadly with utilitarian commitments to creative progress.

Build for Compensation. Exploring ways to adequately compensate people who are contributing to a dataset will be important to preserving incentives to create. While the nature of generative AI makes item‑level attribution challenging, a financial model could be adopted to compensate artists whose work contributes meaningfully to datasets or creative output. At the same time, such tools, consistent with utilitarian principles, could be designed to identify and support fair‑use/fair‑dealing usages of copyrighted material, such as for critical commentary or satire.

Furthermore, in the realm of MIR, ensuring compensation through design will necessitate tools that can accurately identify infringement. Current copyright evaluation systems, such as YouTube’s Content ID and Facebook’s Rights Manager, still lack the nuance needed to detect copyright infringement on platforms; despite longstanding criticism, they continue to misidentify copyrighted material, are poorly equipped to evaluate fair use, and flag works clearly in the public domain (Berkowitz, 2023). More comprehensive approaches for dealing with the increasingly complicated task of identifying AI‑generated work are needed, such as better algorithmic transparency for ID and compensation systems (Berkowitz, 2023), improved metadata/linked‑data management and processing capabilities, (Berkowitz, 2023), more sophisticated feature‑extraction techniques (Cros Vila et al., 2025), and expert human intervention for special or difficult cases (Berkowitz, 2023). Also, supporting other interpretations of compensation, such as gratitude or knowledge sharing (Knearem et al., 2019), may provide additional incentives for creatives who are less interested in economic remuneration.

5.2 Lockean design

Lockean theory could be supportive or circumspect regarding AI tools, depending on how they are designed and implemented. AI could make it easier to expend intellectual labor and create new species of creative property; conversely, it could also facilitate the exploitation of other people’s creations without permission, attribution, or compensation. In addition, the tools could also result in creations with minimal input from human creators, which does not provide strong justification for property ownership. Tools championing the Lockean approach should be built for version control, transparency, and with an eye toward the commons.

Build for Version Control. Version control is a strategy that could acknowledge and document how human efforts factor into the creations of the AI tool. Traceability would make it easier for Lockean‑minded creators to assert legitimacy in rights over creative works and to be able to demonstrate the extent of their contributions. For example, designing to capture the histories of an AI model’s probability statistics (Huang et al., 2025) could support users’ authorship claims by allowing them to better examine how the model’s output corresponds to authorial intent. In addition, clarifying whether and how versioning is carried out in generative contexts—e.g., defining the degree to which a system supports more intuitive creation or stricter record‑keeping, or allowing users to switch between automatic and user‑initiated versioning (Manesh et al., 2024) where versioning control is adaptable to a position in the creative process (Sterman et al., 2022)—permits creators to assert how much granular detail they may need in order to demonstrate a property claim on the completed creative work.

Build for Transparency. A Lockean approach could involve building processes that retain, and make easily accessible, data about AI’s role in ideation and the generation of content throughout (and beyond) the entire life‑cycle of creation. This comports with concepts discussed in human–AI collaboration (Walmsley, 2021; Xu et al., 2019; Zhou et al., 2022), fulfilling the desire of Lockean‑orientated users to maintain control. Transparency in AI creation that clearly shows how AI suggestions were incorporated into the creator’s own ideas would help other creators and audiences understand the boundaries of the creative property that can be rightfully attributed to human ingenuity.

Build for the Commons. Lockean theory, while providing a basis for property rights, also contains caveats against waste or spoilage of resources. Modern Lockeans have sought to balance IP claims with commitments to free speech and a healthy commons of ideas and works open to all (Gordon, 1993). Building out access to freely available intellectual objects within the AI system—for instance, making it easier for creators to access and transform public‑domain or creative‑commons works—can help creatives see value in, and contribute to, the intellectual commons. As AI usage grows, long‑term support for the idea of the commons will require trustworthy labeling and attribution systems, especially for co‑created works (He et al., 2025). This will ensure that authorship for works in the commons can be traced and that AI‑derived elements of works are clear to future generations.

5.3 Personality‑based design

Arguments with implicit personality‑based roots have been some of the strongest critics of runaway AI creation, and it is hard to envision AI systems that would assuage concerns that AI tools will lose the human in the human–AI creation process. However, novice users, disabled users, or creators exploring new artistic genres might recognize the value of AI as a catalyst to bring out their creative voices (Louie et al., 2020). Tools championing this approach should be built for intention, creative control, and idea generation.

Build for Intention. AI tools that help check and confirm the creator’s intention during the creation process would be meaningful for the personality‑based approach. This may be important for novice users who may not have a solid idea of what they want to create, or who lack the domain knowledge to clearly express their intention; in this way, the tool may help them articulate an individual vision in addition to aiding in the expression of that vision. A tool that better retains the information from prior interactions may be able to provide outcomes that are more closely aligned with the creator’s broader vision or direction beyond the creation of a single piece of music (Fu et al., 2025).

In addition, tools supporting intention should be mindful of the ways that artistic goals can fluctuate and adjust over the course of the creative process. Providing tools that allow users to control systems when their creative goals are vaguely defined or flexible (Chung, 2022) and implementing tools that can illuminate the cognitive mechanisms behind creative thinking (Gero et al., 2023; Newman et al., 2023; O’Toole and Horvát, 2024) could better help artists establish, build, or change their authorial voice.

Build for Creative Control. Another viable feature could be the implementation of hard on–off switches and compartmentalization capabilities for how AI is used on a creator’s own works. Creatives who seek a high degree of control may want the ability to turn AI on only for specific, well‑defined usages and to be able to shut AI off when that usage is completed (or if it fails to realize the creative vision). Features that permit specific, bounded, and limited use of AI would afford personality‑based advocates the ability to create on their own terms (Gelineck and Serafin, 2009).

Similarly, creators may wish to shut off certain paths in the creation process rather than simply conducting a series of prompting and selecting. Previous research in MIR has shown that expert users have specific preferences for how AI should be integrated, as well as clear boundaries for where it should not be used (Gero et al., 2023; Krol et al., 2025; Newman et al., 2023). Since much creativity does not follow a linear path, a feature that helps creators better understand and express themselves in non‑linear ways, rather than being confined to predetermined pathways bounded by AI‑suggested samples, would support user control Deruty et al. (2022). Recent work has also surfaced wishes for a more iterative experience in music creation with AI (Choi et al., 2025; Fu et al., 2025), which could be accomplished by supporting easier backtracking in the creation process so that users can revisit past drafts of music pieces and shift the creative course. This process has been shown to be particularly useful in early stages of creative processes, when creators test and prototype ideas through iteration (Fu et al., 2025; Hammad et al., 2025; Newman et al., 2023; Siddiqui et al., 2025).

Build for Idea Generation. Current AI applications are great for creating things fast that are good enough, but not necessarily great for creating things carefully that strive to be profound or excellent. It would be useful to think about crafting the role of AI tools to become partners for idea generation, facilitated by a longer extensive dialogue between the human creator and AI where the creator tries to communicate goals and the AI suggests creative directions (Guo et al., 2024). For those creators who seek it, AI could become closer to a true collaborator or authorial partner rather than simply offering the human a choice of successively refined outcomes.

5.4 Designing across conceptions of IP

Copyright, and IP concepts more generally, are deeply rooted in cultural contexts. Adopting a single approach to design may overlook other important perspectives and practices, and our aim is not to emphasize one to the exclusion of the others. Instead, it is to offer approaches that support an overarching goal of IP practice: to strike a balance between personal creative expression and public creative production. Each philosophy approaches this tension differently, and designers should consider how their systems may favor one side over the other. AI systems that facilitate transitions between IP perspectives can both empower users to navigate these complexities and help foster Merges (2011)’s ‘midlevel agreement’ on IP practice even in the face of competing fundamental principles.

More broadly, we argue that much of the existing research in the AI–MIR space tends to narrowly focus on practical or technical concerns, often at the expense of deeper conceptual engagement, such as critically reflecting on how AI will impact creation from a more human‑centered perspective. In the MIR field, technical development frequently dominates the academic discourse, leaving little room for reflection on the underlying philosophical, legal, and cultural assumptions that influence design decisions. Acknowledging the value of interdisciplinarity and engaging more intentionally with disciplines such as philosophy, law, and cultural studies could help enrich these discussions (Kanhov et al., 2024) and lead to more holistic and reflective system design.

Lastly, we emphasize the importance of studying human–AI interaction beyond short‑term, transactional use. Simply observing users interacting with a system in an experimental setting and collecting immediate feedback does not guarantee meaningful or lasting impact, such as a stronger sense of creative agency or continued motivation to use the system that leads to long‑term adoption. Paying more attention to broader contexts of creation and to the people whose practices are impacted by these tools is important for discerning the real implications of AI systems in creative domains.