How do authorship and creativity apply when an artwork is generated by AI rather than a human artist?

The application of authorship and creativity to AI-generated artwork is a deeply contested and nuanced issue, intersecting philosophy, aesthetics, law, technology, and audience perception. The emergence of advanced generative models has destabilized traditional categories of human authorship and creative agency, compelling both scholars and practitioners to reconsider foundational assumptions about what it means to create and to be credited as an author.

---

1. Authorship: Human Agency versus Machine Generation

Human Agency and the Shifting Role of the "Author"

In traditional frameworks, authorship is synonymous with human agency—intentionality, originality, and personal expression. When an artwork is produced by an AI system, the locus of authorship shifts. The AI does not possess consciousness, intent, or subjective experience; it is a tool, even if a highly sophisticated one. Instead, authorship is often distributed among several human agents: the AI system’s developers and trainers, curators of data, the designers of prompts, and end-users making creative choices during the generation process[1][2]. This division has prompted a reconsideration of single-authorship models and has fueled debates around co-authorship and collaborative creativity with non-human systems[3].

Legal frameworks currently reinforce the primacy of human authorship. For instance, in the US and most other jurisdictions, copyright is only granted to works exhibiting human creative input. Purely AI-generated outputs—where humans play little more than an initiating or evaluative role—are ineligible for copyright protection[1][3]. However, legal systems are increasingly pressed to refine or revise these standards, especially as human involvement in prompting or curating AI outputs becomes more significant[3].

Philosophical and Socio-cultural Foundations

Scholars emphasize that the status and recognition of AI-generated creations also depend on broader philosophical and socio-cultural paradigms. These include views on social subjectivity, cultural value, and social context for evaluating and attributing creative acts[4]. Distinctions are often drawn between AI as a mere tool—invoked by the human’s creative vision—and as a system that potentially contributes meaningful novelty or complexity to the process.

---

2. Creativity: Human, Machine, and Hybrid Forms

Human and Computational Models of Creativity

Creativity in human contexts integrates novelty, value, emotional insight, and cultural awareness. AI systems, by contrast, generate novel combinations by statistically mining large datasets without subjective experience or understanding[1][5]. This procedural or "computational" creativity—sometimes described as emergent or algorithmic—provokes divisive debates. Some theorists argue that AI’s outputs represent a new domain of creativity: distinct from, but influenced by, human mental processes[1][2]. Others maintain that since AI lacks intentionality, its "creativity" is ultimately reducible to the design choices, training datasets, and evaluative interventions of human agents[3][4].

Audience Perception and the "Mind" of AI

Empirical research shows that artworks labeled as "AI-generated" are generally ascribed less "mind"—that is, less agency and experience—compared to those attributed to humans. These perceptions align with lower evaluations of creative worth, highlighting persistent biases about machine agency and creative authenticity[5]. Importantly, the ascription of creative agency is not simply a technical or ontological matter but is bound to social attitudes towards the intentionality and experiential depth presumed necessary for "true" creativity[5][6].

---

3. Legal and Ethical Landscape: Copyright, Credit, and Attribution

Copyright and Originality

Copyright law is an arena where questions of authorship and creativity become highly consequential. The prevailing doctrine in many countries insists on a human origin for copyrightable works, which directly impacts the legitimacy and potential monetization of fully AI-generated art[1][3]. Legal cases such as Li v Liu in China suggest a gradual move towards recognizing nuanced forms of collaboration, evaluating "levels of human creativity and forms of artistic expression" in tandem with AI agency[3]. Nonetheless, there are ongoing disputes about how much creative input from a human is needed for copyright to vest, particularly as AI systems grow more independent in producing outputs[1][3].

Attribution, Credit, and Derivativness

Another ethical challenge is how to attribute credit in AI-generated artworks. Should credit go to the end-user, the developer, the trainer, or be diffused across a collective? Furthermore, since AI systems are trained on vast quantities of existing works, questions regarding the originality and derivative nature of their outputs persist. Some argue that the outputs risk infringing on the rights of original artists, opening a Pandora’s box related to appropriation and cultural exploitation[1][3].

---

4. Artistic and Cultural Perspectives

Practice-Based and Audience-Oriented Considerations

Contemporary artists increasingly treat AI as a medium or co-creative partner, much as previous generations used photography, film, or software[2]. In such contexts, the human artist’s authorship may be asserted through conceptual framing, prompt engineering, curatorial decisions, and post-processing. The meaning and interpretation of AI-generated art is further shaped by the audience’s knowledge of its non-human origins, prompting reevaluation of intentionality, aesthetics, and human exceptionalism[6][7].

Reception, Biases, and Societal Implications

AI-generated art is not immune to the biases embedded in training data or prompt design. Recent research documents, for example, how AI-generated images reproduce problematic cultural tropes, such as age stereotypes—evidencing that the creative process is not value-neutral, and reinforcing the need for critical human oversight and "care-full" entanglement with ethical concerns[8][9].

---

5. Future Directions: Towards Hybrid Authorship Models

As generative AI systems become more capable, the boundaries between human and machine creativity will blur even further[10]. Co-authorship models and revised legal definitions may gain acceptance, alongside new ethical frameworks for assigning credit and responsibility. There is active debate around the development of sui generis rights or AI-specific licensing structures to accommodate the distinctive status of machine-generated works[1][3].

Furthermore, the discourse increasingly recognizes that not only should we ask what AI can do for art, but how artistic practices and critical frameworks might shape the future trajectory of AI itself[11].

---

Conclusion:AI-generated artwork fundamentally disrupts, yet also enriches, our notions of authorship and creativity. While legal and philosophical systems currently insist on a decisive human element, the growing sophistication of generative systems compels ongoing revision and debate. Ultimately, authorship and creativity in the AI context are best understood as distributed, hybrid, and evolving—responses should be tailored to reflect the multiplicity of human inputs, the capacities of AI systems, and the expectations and values of society[1][2][3][4][5].