Culture has always evolved in relation to technology, each transformation reshaping not only how cultural forms are produced, but how they are experienced, shared and valued. What feels different today is not only the speed of change, but the nature of the relationship itself, which seems to be shifting from one of tools and mediation to something closer to infrastructure and condition.
Artificial intelligence is often described as a tool that is beginning to transform cultural production. But this framing no longer seems sufficient. What is unfolding is not simply the adoption of new tools, but a shift in the conditions under which culture is produced, shared and made meaningful.
A recent UNESCO report on artificial intelligence and culture points to both the opportunities and the risks emerging across cultural ecosystems. Yet when these dynamics are observed in practice, across institutions, industries and creative processes, the transformation appears more structural. The question is no longer only what AI can do for culture, but what is happening to culture through AI.
For much of the twentieth century, the possibilities of cultural production were shaped by access to education, to institutions and to networks of recognition. Those conditions have not disappeared, but they are now being reconfigured by another layer of access that is less visible but equally decisive: access to data, to computational capacity and to the platforms where visibility and distribution are negotiated. Creativity, in this sense, is no longer only a human or social process, but one that is increasingly entangled with infrastructures that are neither neutral nor evenly distributed. What can be imagined, produced and circulated depends, to a significant extent, on systems that operate beyond the immediate control of those who use them.
This transformation becomes particularly tangible in music, where AI has moved in a relatively short time from experimentation to everyday use. Generative systems are capable of composing, arranging and producing music in ways that would have required significant resources only a few years ago, and for many artists this opens an expanded creative space. AI can function as a partner in the process, enabling rapid iteration and exploration, and allowing ideas to be tested and developed at a pace that reshapes the rhythm of creative work. At the same time, these systems are trained on vast corpora of existing music, absorbing patterns and structures that have been collectively produced over decades.
The resulting abundance of generated content is inseparable from processes that remain largely opaque, where the boundary between influence, imitation and extraction becomes increasingly difficult to define. Recent industry estimates, including those from CISAC, suggest that generative AI could significantly impact creators’ income streams in the coming years, particularly in music and audiovisual sectors, raising broader questions about how value is constructed and distributed when creation itself becomes easier to generate and harder to attribute.
If music illustrates the scale and speed of change, the performing arts offer a perspective that makes visible both the possibilities and the limits of these technologies. Theatre, dance and live performance are grounded in presence, in the shared experience of time and space between performers and audiences, and this introduces a dimension that is not easily captured by computational systems. Nevertheless, AI is being integrated into these practices through choreographic tools trained on movement archives or performances that incorporate real-time interaction between audience input and algorithmic responses. What emerges in these contexts is not a simple replacement of human agency, but a reconfiguration of it, in which authorship becomes more distributed, the boundaries between creator and spectator more porous, and the performance itself more adaptive. At the same time, these developments bring into sharper focus what remains irreducible: the embodied, situated nature of human experience, and the forms of meaning that arise from it.
A similar, if less visible, transformation is taking place in museums and cultural institutions. Artificial intelligence is increasingly used to restore artefacts, reconstruct heritage sites and expand access to collections, aligning with long-standing institutional goals. Projects such as the digital reconstruction of destroyed heritage sites in Palmyra, Syria, illustrate how AI and 3D modelling can be used to recreate cultural environments that have been partially lost due to conflict. These reconstructions do not simply restore objects; they attempt to rebuild contexts, allowing both researchers and the public to engage with heritage in ways that would otherwise be impossible.
At the same time, these developments point towards a deeper shift in how collections are understood. They are no longer only objects to be preserved, but also datasets to be processed, connected and reinterpreted. This transition redefines the role of cultural institutions, moving them from custodianship towards mediation within complex, data-driven environments. It raises questions that go beyond technical implementation and touch on how knowledge is constructed and represented, including who defines the narratives that emerge from these systems, what forms of cultural expression are made visible, and what is lost when context is translated into data.
Perhaps the most consequential aspect of this transformation lies in a layer that is not immediately apparent. AI systems are not only applied to culture; they are built from it. The texts, images, sounds and everyday digital traces that constitute contemporary cultural life form the material through which these systems are trained and acquire their capacity to generate meaning. Culture, in this sense, is not simply an object of AI, but a condition of its possibility. Yet this collective dimension remains largely unrecognised within existing frameworks. Cultural data is rarely treated as a shared resource requiring governance and protection, and the value generated through its use is unevenly distributed. The result is a structural asymmetry in which cultural production feeds AI systems at scale, while the benefits that emerge from these systems do not necessarily return to the communities that sustain that production.
Understanding these developments as a simple case of technological disruption would be misleading. What is at stake is the reconfiguration of the systems through which culture is produced and valued. Across different contexts, technological capabilities are advancing at a pace that exceeds the capacity of institutional, regulatory and educational frameworks to respond, creating a growing gap between what is technically possible and what is socially and culturally governed. This gap is not only a source of risk, but also a space in which new forms of organisation and collaboration can emerge.
Reframing the conversation therefore becomes necessary. Rather than asking what AI can do for culture, it may be more useful to ask what kind of cultural systems are being constructed through AI. This shift brings into focus the relationships between infrastructure, governance and value, and invites a more deliberate engagement with the consequences of technological development. It suggests that the central challenge is not simply to adopt or resist these technologies, but to shape the conditions under which they operate.
From this perspective, the transformation of culture through AI is neither singular nor predetermined. It consists of a series of interconnected shifts that challenge established assumptions about creativity, authorship and value, and that unfold across different domains in uneven ways. Engaging with these shifts requires moving beyond simplified narratives of disruption or progress and towards a more nuanced understanding of the systems in which cultural life is embedded.
Culture has always evolved in relation to technology, but what defines the present moment is the extent to which that relationship is being reorganised at the level of infrastructure. Artificial intelligence does not simply introduce new possibilities for creation; it reshapes the conditions under which those possibilities emerge, distributing agency, visibility and value in new ways. What is at stake, therefore, is not only how culture adapts to AI, but how the systems we design around these technologies will shape what culture can become.
References and further reading
– Buolamwini, J., & Gebru, T. (2018). Gender shades: Intersectional accuracy disparities in commercial gender classification. Proceedings of Machine Learning Research, 81, 1–15.
– UNESCO. (2025). Report of the Independent Expert Group on Artificial Intelligence and Culture. https://www.unesco.org/sites/default/files/medias/fichiers/2025/09/CULTAI_Report%20of%20the%20Independent%20Expert%20Group%20on%20Artificial%20Intelligence%20and%20Culture%20%28final%20online%20version%29%201.pdf
– CISAC. (2024). Study on the economic impact of generative AI in the music and audiovisual industries.
– UNESCO. (2021). Recommendation on the ethics of artificial intelligence.
Manel González-Piñero is a professor of economics and innovation at the University of Barcelona and ESMUC, focusing on cultural and creative industries.