Last year, I wrote a piece on how platform power in the music industry has systematically devalued the labor of musicians and pushed them into precarity. In 2025, I was at the Internet Governance Forum – an annual multistakeholder platform to discuss emerging issues on internet governance. At last year’s gathering, voicing concerns about the rise of generative AI tools in entertainment and culture, actor Joseph Gordon-Levitt said, “the data that humans produce — our writings and our voices and our connections, our experiences, our ideas — should belong to us. And that any economic value that’s generated from this data should be shared with the humans that produce it.”
That simple premise is increasingly a pipe dream in today’s creative industries.
The AI art heist
Over the last three years or so, musicians, actors, artists, and other cultural workers have been confronted with a startling and unsettling realization – the very place that promised them visibility and discovery (the internet) instead served up their life’s work on a silver platter for AI extraction pipelines such as crawlers and scrapers. Illustrators across platforms such as ArtStation and Instagram began to observe images they had produced over the years reappearing in AI-generated content online. LAOIN-5B (a dataset used to train these generative AI tools) has scraped a whopping 5.3 billion images from the open internet. Online literature platforms that functioned as a place for emerging writers to test their chops have become flooded with AI-authored texts, at a rate far outpacing what humans can realistically produce. Spotify is pumping its playlists full of cheap music made by generative AI models that often go on to become hits, thereby also replacing the precarious ghost artists who were previously churning out songs for Spotify’s curated playlists.
The response has been varied. First, a wave of anger has arisen to meet these developments, to protest AI’s theft of original work, rendering the very artists responsible for creating it “obsolete”. Second, there is widespread anxiety about what this means for our relationship as a species with art and cultural goods, prompting philosophical ruminations on the question of authorship, artistic intent, and collective meaning-making. The New Yorker ran a piece titled “Is It Wrong to Write a Book with A.I.?”, prompting a frenzied condemnation in the comments section. In this way, the “man v. machine” framing has once again resurrected the relevance of these age-old questions.
There is an overall fixation with the question of labor displacement as a result of AI (generative AI in particular), and indeed, these anxieties are very real.
There is an overall fixation with the question of labor displacement as a result of AI (generative AI in particular), and indeed, these anxieties are very real. Voice actors are being asked to sign away rights to their voices so that they can be used to generate synthetic replicas through generative AI. There are widespread fears that the entire influencer economy could be replaced by virtual clones and personalities. In 2023, a viral AI-generated track mimicking Drake and The Weeknd was removed from streaming platforms after rights disputes, showing that even industry behemoths are not spared, even if they have more effective avenues to resist and protest as a result of their power.
Yet, this obsession with labor replacement misses two critical pieces of the conversation. First, what does it mean for working conditions in the creative industry as a whole, and where does this figure in digital capitalism’s project of precarity? Second, what does the reduction of artistic intent to computation tell us about the process of meaning-making?
Yet, this obsession with labor replacement misses two critical pieces of the conversation.
Although these are currently two parallel conversations, my argument in this piece is that they should be seen as two sides of the same coin.
Techno-optimists argue that artists will simply have to learn to use AI, or risk being left behind. They compare this to any other innovation, like the synthesiser, autotune, typewriter, or even spell-check.
It is very much technically possible that artists can create in partnership with generative AI tools, in a manner that pushes the boundaries of form and substance both. What is clear, however, is that in the current political economy of AI, this emancipatory future for artists will not materialize. The machine is built on theft and appropriation, whose bottom line is only profit.
What policy discourse is missing
Generative AI systems are trained on a vast corpus of cultural material scraped from the open internet, collapsing the distinction between fair use and flat-out theft. The debate has hinged very much on intellectual property regimes, and copyright in particular. There have been a slew of recent judgments on this issue – in some cases, courts have ruled that scraping shadow libraries to create training datasets is a “transformative use”, and not a reproduction, thereby falling within the fair use exemption. In others, where AI directly returned copies of the training material as outputs, courts have ruled these constitute a violation of copyright. In any case, copyright is increasingly becoming the battleground in which the future of our cultural commons is being fought.
It is worth remembering here that copyright, like patents and trademarks, very rarely rests with individual artists and workers.
We are not likely to win this fight here. It is worth remembering here that copyright, like patents and trademarks, very rarely rests with individual artists and workers. Instead, it is corporations that own them – film studios, record labels, publishers, etc. In fact, intellectual property regimes are moored in a history of individual property ownership and enclosure of the commons. Today, the mechanics of this legal regime are being used to sacrifice cultural workers and their labor at the altar of AI. Spotify has entered into licensing deals with Universal Music Group, Sony Music, and Warner Music (who collectively control nearly all recorded music) to use their databases to train AI tools, with a provision for artist compensation (which is nominal at best, compared to the millions Spotify seeks to gain). Lionsgate Studios has partnered with an AI firm, Runway, allowing it to train its database on its film and voice recordings. Overall, the preoccupation with intellectual property suggests that our cultural commons is up for grabs – under certain conditions. None of this has anything to do with protecting cultural workers or our cultural rights.
What if, instead of intellectual property, we locate this discussion in the rights of creators as essential workers?
Cultural work as social infrastructure
Much has been written about the flattening of our cultural sphere through algorithmic curation, which simultaneously feels personalized and formulaic. Given that generative AI is essentially a technology of prediction (predicting the next token through averages), the real risk that we encounter is not that AI will write a bad book, thereby reinvigorating the livelihood of all ghostwriters and freelancers. It is that AI is capable of writing a perfectly competent average book, yet one that is completely devoid of intention and agency.
Genuine creativity requires “intentional agency”.
It is becoming very clear that AI will not be capable of authenticity – that is, the ability to create from a unified subjective experience. Whether it is an influencer integrating brand partnerships into daily vlogs, an essayist chronicling her youth, or an artist’s freehand sketches, all of these outputs are saturated with artistic choices of “self-disclosure” (what is revealed to the world). Genuine creativity requires “intentional agency”. This is precisely the aperture that AI slams shut – the shared process of meaning-making through what is shown to us and what we see. Unfortunately, it is precisely this illusion of authentic meaning that AI is well-equipped to replicate. In the process, what we also lose is our skill of cultural discernment.
The reduction of art to computation is thus not just a technological shift – it is a wholesale reconfiguration of how societies produce shared narratives.
Our cultural commons shapes how societies understand themselves, how we process conflict, and how we imagine futures. The labor that sustains this commons is foundational, and not peripheral. Existing policy frameworks still treat art as an economic sector rather than as social infrastructure. What is often missing from policy and economic debates, therefore, is a recognition that artistic and aesthetic intent is itself a form of essential labor that is necessary for our collective political and civic health.
Existing policy frameworks still treat art as an economic sector rather than as social infrastructure. What is often missing from policy and economic debates, therefore, is a recognition that artistic and aesthetic intent is itself a form of essential labor that is necessary for our collective political and civic health.
When this becomes the starting point, intellectual property regimes appear as an obstacle rather than a solution. Through a worker-centric approach that puts individual and collective rights at the centre, we can steer policy frameworks towards creating a fertile ground for a healthy commons ecosystem. Three things are central to this effort:
- i) Reforming the intellectual property regime to give artists, communities, and creators greater control over how their artistic outputs are represented and used within AI systems. Against the backdrop of copyright, alternative licensing systems have emerged, particularly to counter the hegemony of Western digital capitalism and safeguard indigenous knowledge systems from AI’s appropriation.
- ii) Creating channels for equitable value redistribution that recognize the provenance, labor, and affective work of artists and creators. Models of individual opt-out and compensation are ill-equipped to tackle the structural overhaul of culture precipitated by AI.
iii) Protect artists and cultural workers from precarity by undertaking an overhaul of working conditions within the creative sector.
As Audre Lorde wrote, “Poetry is not a luxury. It is a vital necessity of our existence. It forms the quality of the light within which we predicate our hopes and dreams toward survival and change, first made into language, then into idea, then into more tangible action.”
My provocation is thus simple: a healthy cultural commons is a public good, and thereby, cultural workers are essential workers whose rights and dignity are fast unravelling in the age of AI.
