Introduction

How can we use Generative AI (GenAI) to see differently and imagine different political cultures created by a combination of workers’ movements and hip-hop? Do artistic practices embedded within provincial cultures entail strategies when dealing with the Anglo-American cultural biases of GenAI, and how can the provinces decenter the realisms generated by GenAI? The ways that current GenAI platforms imitate human language and photo realism can be seen in line with how a naïve realism has been a central aesthetic in the design of computer interfaces, as discussed further below. Ørum addresses this together with the history of how Danish politics, as in other similar countries, have turned towards marketization, which will also be discussed further. #freedomlibertyandhiphop is a series of artistic projects and exhibitions through which Kristoffer Ørum reconsiders recent history and imagines alternatives. Overall, it portrays a political, provincial utopia, which is still recognizable, though also filled with co-created AI narrative.

As an artist, Kristoffer Ørum uses digital media to question current realities and contemplate alternate possibilities, addressing themes like climate change (Djurs Archipelago and Ebeltoft After the Flood) invasive species (Signal Crayfish) and collectively imagined monsters (The Kolding Creature and Mundane Monsters). Recently, he has used GenAI to imagine and project alternative histories. He has related these alternative histories to various local towns in Denmark and abroad, like Hvidovre, Aalborg, Malmö, Riga, and Aarhus, in exhibitions in these places.

Ørum was born in the Copenhagen suburb Farum in 1975 and has described how he was inspired by, and part of, early Danish hip-hop and graffiti culture. Like other suburbs from the late 1960s and 1970s, Farum was characterized by grey concrete blocks conceived by welfare politicians as modern, comfortable living that instead turned into drab neighborhoods. Through graffiti and hip-hop Ørum envisioned a more colorful world where things could be changed, even though he did not have access to the American scene besides a couple of shared cassette tapes. Farum was far from the center of things in a boring suburb of Copenhagen. This background led Ørum to imagine a connection between the Nordic workers’ movement and early hip-hop, including what would have happened if this never-realized connection had materialized. He is especially interested in whether cultural, political and utopian alternative possibilities remain hidden in this imagined connection (Nyeng, 2023; Ørum, 2025). Furthermore, he uses this narrative frame to challenge GenAI image platforms to get beyond its naïve (photo) realism towards a more complex, critically reflective mode of realism, that will be discussed below.

In #freedomlibertyandhiphop [#frihedlighedoghiphop] Ørum artistically employs GenAI images generated via Stable Diffusion to explore what might have happened to the Danish worker’s movement and late 20th-century culture if hip-hop culture had been adopted and taken seriously in Denmark. The project emerged on Ørum’s Instagram account in 2023 as a series of numbered studies, some of which were exhibited at the Sydhavn Station Gallery in Copenhagen in 2023. As a series, it has carried over into later exhibitions like Mr Spray, Monuments of A Fictitious Past (Riga) Aalborg is the Future (Aalborg), Hvidovre Makes Good Times Better (Hvidovre), and culminating with Aarhus is #1 (Aarhus) in 2025. His work has also been part of group exhibitions in Denmark and internationally in Sweden and the USA.

Besides portraying a counterfactual history, #freedomlibertyandhiphop explores the realist modes of GenAI by depicting an alternative reality that conserves marks of GenAI image-making in the early-mid 2020s such as distorted backgrounds, architecture, and landscapes. Ørum employs GenAI as a co-creative partner to depict a slightly surreal artificial version of Danish culture and politics, highlighting the types of depictions generated by GenAI platforms. By provoking and countering the built-in tendencies of GenAI toward a form of Anglo-American monoculturalism (Rettberg, 2023), Ørum unlocks new possibilities for critique, and GenAI is transformed into a tool for reflecting on alternative politics and realities. Beyond its humor and critical reflection on Denmark’s recent history, #freedomlibertyandhiphop hints at how we can design and think about GenAI in the future—as a creative and reflective tool. The images are published in a thematic series on Instagram, accompanied by texts written by Ørum that anchor, describe, and narrate the pictures. The texts are (human-)written by Ørum and are inspired by the narrative prompting and the generated images, including the oversized record players, multiuser spray cans and other possible glitches generated from the data sets in co-creating the alternative history.

Hip-Hop Welfare

From 1929 to 1981 Denmark developed into a Nordic welfare state with social democratic governments and strong labor unions. Under the #freedomlibertyandhiphop hashtag, Ørum attempts to imagine an alternative political and cultural history from the period of social democratic Danish Prime Ministers Anker Jørgensen (1972 – 81), who resigned because of a large public deficit, through to Poul Nyrup Rasmussen (1993 – 2001), and Helle Thorning Schmidt (2011 – 15). Since 1981, right-wing governments have periodically replaced social democratic rule, introducing neoliberal and national conservative ideologies that have partly supplanted the original values of the workers’ movement. The original values of freedom, liberty and equality were gradually suppressed by austerity measures, neoliberal market thinking and an increasing focus on ethno-national identity. Even Social Democrats were influenced by the ‘Third Way’ ideas of the competition state embraced by Anthony Giddens and Tony Blair (Green-Pedersen and van Kersbergen, 2002; Cerny, 1997). This also happened beyond Denmark across the Global North. In #freedomlibertyandhiphop, Ørum reroutes this history via hip-hop.

#freedomlibertyandhiphop imagines that the labor movement and social democracy embraced the early Danish hip-hop culture with its distinct do-it-yourself ethos and local roots, rather than turning towards the Third Way and national identity politics after the left-wing took over through the 1980s and again in the 2000s. As mentioned above, this counterfactual imaginary came to Ørum from growing up in the suburbs during the 1980s austerity politics. Consequently, Ørum imagines a political and cultural history that never happened, as Danish hip-hop and the workers’ movement did not, in reality, develop strong mutual connections. However, in Ørum’s alternative history, Danish Prime Minister Anker Jørgensen travelled to the Bronx in May 1974 ‘to study the earliest form of graffiti, break dancing, and DJing’ (as stated by the subtext to Figure 1). Jørgensen became dedicated to the cause, and his trip ‘helped to introduce hip-hop culture to the general public of Denmark, where it has since become a popular and influential form of expression’ (Study number 38, Ørum, 2023-). Thus, instead of accepting the austerity politics in 1981, Ørum retrospectively imagines a new cultural and political blossoming for workers’ culture, inspired by hip-hop.

Figure 1: “Anker Jørgensen” in NYC, Kristoffer Ørum #freedomlibertyandhiphop, study number 38. Reproduced with permission of Kristoffer Ørum.

Danish hip-hop culture emerged in a context where Bronx and Compton were far away places. The internet’s easy access to foreign culture, music, and the right Adidas sneakers was still not established. However, unique Danish provincial expressions from the suburbs and depths of the country gradually emerged. While the USA saw violent rap battles between West Coast and East Coast artists, in Denmark it was somewhat more peaceful and humorous when Østkyst Hustlers [Eastcoast Hustlers] travelled to Jutland and complained, ‘Why is everything so uncool in Jutland?’ [‘Hvorfor er det hele bare så ufedt i Jylland?’] (Østkyst Hustlers, 1995).

Østkyst Hustlers were instrumental in developing Danish rap by making what originated as a parody of American rap and rappers. Subsequently, early Danish rap and hip-hop developed a strong local dimension. This was initially mainly driven by white youth but gradually through the 2000s integrated aspects of non-white Danish culture. Acts such as Outlandish from the suburbs of Copenhagen introduced samples from Arabic and Pakistani music, and the Aarhus V rap scene from Aarhus’ western suburbs – which is also reflected in the exhibition Aarhus is #1 – was also dominated by non-white Danes (Ørum, 2025). Today, Danish rap is a multicultural scene in a still rather homogenous Danish culture, and it is fair to say that this diversity has developed from an early scene that included provincial perspectives and dialects, and now incorporates multicultural Danish idioms.

In the #freedomlibertyandhiphop narrative, this local grounding quickly becomes an integral part of Danish hip-hop’s expression, translating American culture into Danish streets and countryscapes.

Provincially Generated Utopia

Ørum’s use of technology is generally driven by a critical perspective that also affects his artistic expression. His criticism is, however, combined with a playful curiosity that drives him toward exploring alternatives. Rather than flashy hi-tech, he uses technologies that bear the marks of duct tape, recycled computers, and everyday life—albeit in humorously wry versions that point to alternative utopias and dystopias.

In #freedomlibertyandhiphop, he uses the open text-to-image model, Stable Diffusion (mainly XL released in July 2023 and Flux, released by Black Forest in August 2024). Stable Diffusion is, as its name suggests, a diffusion model trained with the objective of removing successive applications of noise on training images. It is trained on pairs of images and captions taken from the LAION-5B publicly available dataset, derived from Common Crawl data scraped from the web. Besides openly sharing its source code and model, this model stands out for its ability to be downloaded and run on ordinary computers with probably less energy costs and carbon emissions than cloud-based models, and with the possibility for users to use renewable green electricity (Utz and DiPaola, 2023; Weinbach, 2024). However, as with other models, it is trained on web-data which disproportionally contains American data, resulting in an Anglo-American cultural bias. In 2023 it also made errors with human limbs, strange spelling, and most pictures had blurred backgrounds due to the limited size of 3.5 million parameters (Wikipedia contributors, 2025). Finally, it is easier to get around its fine-tuning and use it for images picturing people or things censored on other GenAI platforms. This makes it useful for art, including for picturing Danish politicians besides, of course, for fake images and pornography.

Significant effort is consequently required to make the images appear authentically Danish, given the global/American bias inherent in the model’s dataset. For example, pictures of churches that do not resemble Danish white chalked Romanesque stone churches pop up, and other aspects of architecture or landscape also fail to look Danish. Ørum himself explains that he often must prompt the scenes in the images to take place in Montana, as the model lacks sufficient Danish data and Montana is the closest approximation to a Danish landscape. However, the problem often arises when churches and barns end up looking American, manifesting the cultural bias in the training data. Still, the cultural bias goes deeper than the look of the landscape and buildings. As Scott Rettberg points out in relation to language models like ChatGPT, even when ‘learning other tongues, it is learning to speak American in the cultural perspectives it represents’ (Rettberg, 2023: n.p.). Besides the landscapes, barns and churches, this deeper cultural bias is also true with image models and Ørum works explicitly with this bias.

In study number 35, Ørum imagines a ‘Djurs Renaissance’ within Danish hip-hop, which, according to the accompanying text, led to rappers starting to use the local Djursland dialect—a trend that spread internationally. Words like ‘nølle,’ ‘hæster’n,’ and ‘Wårhær’ become mainstream in hip-hop after the Djurs Renaissance, which also incorporates natural sounds and folk music. However, these words are, like the images, not typical of Djurslandic dialects but rather a sonic approximation. In study number 35 (Figure 2), the images depict people meeting with their colorful gramophones among beets and carrots in Djurslandic mud and rain, rapping with North Djursland’s cliffs (or an adapted American cliff) in the background, and DJing for a squirrel in a remote forest.

Figure 2: The Djurs Renaissance, Kristoffer Ørum: #freedomlibertyandhiphop, study number 35. Reproduced with permission of Kristoffer Ørum.

For people who grew up in Djursland (like the author) or in a similar remote countryside, these pictures are similarly recognizable and incisive in their portrayal of provincial life and culture. They are recognizable in the many, sometimes quite challenging, attempts to make things happen far from the big cities. As an organizer or performer, one could often feel a bit like the DJ with the squirrel, showing one’s (lack of) talent to nobody, dreaming to get away. As a lonely participant, there might also be a certain lonely squirrel-like feeling, showing up at an event somewhere outside of city limits as one of the few. However, it also points to a specific quality of provincial life and culture and how the provincial can be a specific creative source.

Provincial life offers moments of accomplishment, where a unique atmosphere and joy of creating something independently is combined with a commitment raised by the fact that there are no alternatives within 50 kilometers. A similar provincial feeling even spread in suburbs to larger cities such as the western suburbs of Aarhus and Copenhagen, where people felt socially, culturally or ethnically marginalized. This, in fact, led to the development of hip-hop scenes as argued above, including Ørum’s descriptions of Farum.

A small Nordic country like Denmark even feels provincial faced with global trends from the Anglo-American world. In study number 32 (Figure 3), the accompanying text describes an imaginary scenario where American hip-hop in the 1980s had a significant influence on the Danish welfare state. However, due to limited direct contact, Danish hip-hop developed based on local traditions and creative misunderstandings. #freedomlibertyandhiphop depicts a time before the widespread use of the internet when people in Denmark had much less access to international music, culture, and fashion than today, where even the remotest areas have access to streaming platforms and web shops. Even if streaming platforms like Spotify contain profiling algorithms and playlists that serve to narrow diversity and support the most popular artists (Pelly, 2025), streaming still makes a lot more music (and other cultural content) available everywhere.

Figure 3: Danish translations of American hip-hop, Kristoffer Ørum: #freedomlibertyandhiphop, study number 32. Reproduced with permission of Kristoffer Ørum.

In the 1980s and into the 1990s, the provinces were comparably more marginal before the popularity of computer networks and the web distributed music and culture to the most remote villages (Bello and Garcia, 2021). There were a few national radio and TV programs like the ones hosted by Kim Schumacher, who was renowned in Denmark and often picked up music in New York and spread it back in Denmark (Lund and Christensen, 2020; Modler, 2024). However, it took almost all of the 1980s before the first serious hip-hop album in Danish emerged, MC Einar: The New Style [Den nye stil], in 1988 after several attempts at ‘sloppy plagiarism’ by ‘revue goofballs’ (Olsen). There was considerable creativity in the forced provincialism to cultivate distinct cultures, expressions, and values. Danish hip-hop sounded different from the international Anglo-American scene.

#freedomlibertyandhiphop imagines an alternate timeline where global culture developed differently than today. Global accessibility has changed provincial culture for good and bad from the 1980/90s to today, though sheer online accessibility does not make the change alone nor bring an end to the importance of environment, space and place. Perhaps the kinds of marginalization have changed and increasing economic inequality and xenophobia plays a major role globally and in Denmark. Access to Spotify, Netflix and Facebook has not created equality between the city and the provinces, whether they are in the countryside and suffer from lack of economic development or in the suburb and suffer from xenophobic politics. As mentioned above, hip-hop has been important for addressing the latter.

‘Djurs Renaissance’ is a good example of how Ørum works with the local context in a realistic scenography. The local context is weaved into a slightly anamorphic national and international history that besides the odd charm of the provincial, points to utopian possibilities amidst the mud and beet fields. Out of this comes elements of an alternative history that serves as a corrective to more recent Danish history and culture, which were shaped by national conservative values during the governments of the neo-liberal prime minister Anders Fogh Rasmussen in the 2000s. Meanwhile, the labor movement and the Social Democrats often lacked visionary new ideas for developing the welfare state, had lost sight of innovative culture and art and was giving in to the Third Way. However, this changed, according to Ørum’s narrative, with the world’s first hip-hop democracy, a central part of his alternative history.

The World’s First Hip-Hop Democracy

In the Autumn of 1990, the Social Democrat Helle Thorning-Schmidt opened the Danish parliament (Folketinget) as the first female prime minister, according to Ørum’s accompanying text. She performed a two-hour-long DJ set, emphasizing cultural diversity and promising to make Denmark a more creative and equitable society as the ‘world’s first hip-hop democracy.’ As any follower of Danish politics will know, this is quite a hopeful alternative history. Helle Thorning-Schmidt’s government did not come to power before 2011 and it was curtailed by the Social Liberal Party’s (Radikale Venstre, one of the coalition parties of her government) insistence on continuing the neo-liberal economic policy of the former right-wing government. Thorning-Schmidt’s government furthermore broke down when the other coalition party, The Socialist People’s Party left out of disappointment. The two-party government continued for another year and was not re-elected.

Consequently, the actual Thorning-Schmidt government can be seen as an illustration of the lack of new ideas and innovative culture that #freedomlibertyandhiphop presents an alternative to. The lack can be seen in how the images do not quite resemble their described motives. Copenhagen and the architecture of the parliament (Christiansborg) appear disfigured in the background of several of the images (like Figure 4, right). In some of the pictures (Figure 4, right), Thorning-Schmidt almost resembles herself, while the men in the background’s attire looks too formal and conservative for most Danish members of parliament. In others (Figure 4, left), Thorning-Schmidt looks more like Margaret Thatcher or perhaps Angela Merkel. Though she was the first Danish female prime minister, the left picture in this way might be seen as a haphazard illustration of how she was held back by the economic policies inherited from more conservative colleagues as her government became a victim of Third Way politics. Just as the image set lacks representation of Danish landscapes and buildings, Danish politicians are similarly scarce. Ørum must rely on descriptions in the prompt, generated from AI image analysis tools, to approximate their appearance, resulting in a quality that straddles photorealism from the image sets and prompted corrections. Figure 4 more generally illustrates how the hip-hop renaissance points to an alternative to the necessity-driven neo-liberal politics, emphasizing creativity over the economies of labor supply. However, it also illustrates, at least in a figurative interpretation, how conservative and neo-liberal values penetrate the images of GenAI from behind and within – as background and bias in the image sets.

Figure 4: Helle Thorning-Schmidt opening the Danish Folketing with a DJ-set, Kristoffer Ørum: #freedomlibertyandhiphop, study number 25. Reproduced with permission of Kristoffer Ørum.

In #freedomlibertyandhiphop Ørum imagines that local Danish and Nordic ideas about the welfare state, social democracy, the labor movement, cooperative movements, and participatory democracy seized the opportunity, evolved, and embraced progressive do-it-yourself traditions during the early days of hip-hop. Instead, cultural stagnation occurred, with mainstream pop compilations playing in the background, and right-wing values and concepts of Danish identity taking hold.

As Ørum shows and narrates, Danish hip-hop, as in the USA, gradually becomes corrupted by money and turns into the ‘bling-bling era of Danish hip-hop marked by privatization, crackdowns, and violence’ and as well as becoming ‘associated with commerce and luxury’ (Study number 40, 2023-). There are consequently no heroes, not even in the alternative history. In Denmark hip-hop was influenced by bragging and toxic masculinity, though study number 40 (Figure 5) is also a satire on US hip-hop. In general, #freedomlibertyandhiphop calls for a global and open perspective that doesn’t forget or erase the qualities of the local and our inner provinces but develops them into new visions. It is as if architects would collaborate with residents to create social housing as portrayed in The National Association for Colorful Building Traditions [Landsforeningen Farverig Byggeskik] (Figure 6) instead of building suburban non-places in grey concrete as in Farum and glass and steel corporate office architecture. It is a utopian dream of an alternative history where hip-hop developed a Nordic style that had a political impact and changed cultural history, including architecture. It did not happen, neither in Nordic hip-hop nor politics, but the dream looks enticing, as if it could have been real.

Figure 5: The bling-bling era of Danish hip-hop. Kristoffer Ørum: #freedomlibertyandhiphop, study number 40. Reproduced with permission of Kristoffer Ørum.

Figure 6: The National Association for Colorful Building Traditions. Kristoffer Ørum: #freedomlibertyandhiphop, study number 36. Reproduced with permission of Kristoffer Ørum.

GenAI Realism

Besides a comment on Danish political and cultural history, #freedomlibertyandhiphop is also a comment on and exploration of GenAI’s modes of imaging. Realism has outside the fields of literature and art often been a central aesthetic in the development and design of computer interfaces. It has repeatedly been argued that the interface, if not the entire computer, should be hidden, and transparent in the traditions of GUI, VR, XR, and ubiquitous computing (Pold, 2019). Simultaneously, computing seems increasingly removed from merely representing the object world, moving towards virtual reality in different ways. AI seems to have taken a further step away from representing an external reality. It has been argued repeatedly, that we cannot trust the output of GenAI, that it produces fake images and texts, that we have little way of getting insight into how a specific output is produced by its internal neural networks, and that business logic even prohibits insights into what has been harvested for the data sets and Large Language Models (LLM) (Crawford and Paglen, 2021).

It might consequently be argued that GenAI is less realistic than earlier interfaces and that its output through generative production processes becomes further removed from representing and referring to the object world. However, if realism is not seen as a naïve belief in representation and mimetic imitation—as in much corporate interface design—but rather as a reflection on how reality is produced in and through media technologies and interfaces (Pold, 2005), GenAI images can also be seen as a new more complex development in realism. Already with platformed, ubiquitous, networked and datafied metainterfaces, it has been argued that on top of the illusionism of traditional interfaces, a data realism is added. Metainterfaces are not just a seamless representation of virtual realities but also augmented ones. Consequently, we move beyond the windows metaphor towards a metainterface as a networked enhancement of reality with layers of data (Andersen and Pold, 2018: 36). This development reflects that contemporary realities become increasingly generated by data and through the ‘production of prediction’ of, for example, economic forecasting, climate change models, social metrics, and platformed publics (Mackenzie, 2015). GenAI interfaces can be seen as in line with this tendency, producing output that rather than referring to an object world are ‘data hypothesis’ produced from image sets and LLMs. As Leif Weatherby suggests:

[…] to the extent that we live in a machine-classified world, we contend with the ‘data hypothesis’ as a semiotic surround, a digital environment of semiological abstraction, rife with error and bias but nonetheless all too real as a remediation of our understanding of the world—because it is part of that world (Weatherby, 2025: 105).

Besides the development from the illusionism of virtual windows to the metainterface’s augmented reality of networked data realism, a gradual shift from the metaphor of interaction to profiling and mimesis can be observed. Malthe Stavning Erslev explores and develops Walter Benjamin’s concept of the mimetic faculty and nonsensuous similarity as a mimesis that is more than imitative:

The central insight is that in our appreciation of language, the mimetic faculty situates an aesthetic experience of the sign that gives rise to a surplus value that exceeds (and in a sense mediates) the meaning of it. This is a different understanding of mimetic language than the more conventional use of the term as referring to realism. Instead of being mimetic in the sense that the meaning corresponds to our understanding of reality, the concept of nonsensuous similarities in language asserts that before we can even make such a judgment, the mimetic faculty has already been in play in our very appreciation of the signs themselves (Erslev, 2024: 54).

Erslev consequently argues that the semiotic and mediatic structures should not be overlooked but are part of mimesis, including the surplus value the media produce. Exploring GenAI images or text also becomes a way of exploring their signs and semiotic structure in a kind of meta-semiotic reading that is also exemplified by the meta-prompting Ørum must develop to get something that looks like Danish landscapes, architectures and people. Consequently, this understanding of realism is far more complex than traditional understandings of realism and points to a new kind of GenAI realism, which, however, still mimics human language.

Even if images are a different aesthetic modality than text, GenAI images also follow a realistic style, aiming to mimic other styles of images and ‘ways of seeing’ (Berger, 2008; Azar et al., 2021). In general, GenAI realism can be characterized as skeuomorphic where the new materiality of image/language models and image/text generation procedures is hidden behind an emulation of well-known visual/textual genres to highlight immediate recognition and visual/verbal user-friendliness. Visual GenAI models often simulate existing visual aesthetics including photorealism and can be seen as post-photographic (Malevé and Sluis, 2023). The process of prompting GenAI aims to give users what they want. GenAI emulates the kind of relation to reality of a human, aiming to produce knowledge of and represent the world as if it were an embodied human perspective. However, a GenAI only ‘knows’ and ‘sees’ what it can correlate from its LLM and datasets, and its output language has no direct relation to the object world or any embodied physical presence. Consequently, rather than just simulating ways of seeing, GenAI as ‘data hypothesis’ simulates ways of picturing, including the medialities of images, even though its internal processes are entirely different from cameras and painterly perspective.

In what sense does this relate to #freedomlibertyandhiphop? Kristoffer Ørum has jokingly said that in the beginning of the project he was contacted by some people from Italy asking if the local trains were really painted with abstract squares, as illustrated in Study number 22 (Ørum, 2026). While this was in August 2023 when people were still less acquainted to GenAI images, it points to Ørum’s strategy of navigating within the space of GenAI realism exploring alternative histories that meanwhile seem realistic in #freedomlibertyandhiphop. To be precise, his work explores a GenAI realism in the sense sketched out above. Rather than discussing whether GenAI is realistic in the sense of faithfully representing the object world, the work explores how GenAI is part of and a reflection on a contemporary production of realities driven by neo-liberal and algorithmic capitalism. Ørum essentially uses GenAI to explore how it would look if neo-liberalism, Third Way capitalism, and the competition state had not become dominant ideologies, or to stay with Ørum’s own examples, if Anker Jørgensen’s Social Democrats had not resigned and left the government in 1981 but had renewed their culture and ideas, as envisioned by Ørum (see Figure 1).

Ørum’s specific turn on GenAI realism relates also to his use of GenAI and his provincial approach. Ørum produces the visual language and narrative of #freedomlibertyandhiphop through a critical reflection on its GenAI model. The work explores a ‘tendency’ in its use of GenAI technology. Tendency is a critical concept developed by Walter Benjamin in his discussion of how authors should understand themselves as producers enwrapped in a (re-)production technology that they should seek to critically redesign. Benjamin uses it as a way to develop Marxist media aesthetics, describing the relations between base and superstructure, (re-)production technology and artwork, as productive rather than just an attitude (Benjamin, 1996; Pold, 2018). In this sense, the concept of tendency can be seen in continuation of Benjamin’s thinking of mimesis as including the material mediality, as argued by Erslev above. For a work to have a critical tendency, it should reflect on and redesign its reproduction technology, including GenAI.

Ørum is using GenAI in his co-writing and narrative with a critical tendency by challenging bias through the production of alternative and provincial histories and similarly challenging GenAI by DIY methods rendering work that is different from the hi-tech magic of some GenAI discourses. FreedomEqualityHiphop opens up the possibility of a different understanding of GenAI, and the potential different ways of using it. The humor of the work is often attached to elements where GenAI fails in its illusionistic realism, and we as users realize that this is not the real history, but one that could have been. For a Dane who knows about the actual history, it is obvious that Anker Jørgensen did not discover hip-hop in New York and Helle Thorning Schmidt did not open the parliament with a DJ-set (Figure 1 and Figure 4). But even for someone less versed in newer Danish history there is something twisted about the images, even with Denmark’s history of welfare, social democracy and co-ops.

#freedomlibertyandhiphop imagines a Denmark that could have been if Denmark, like other countries, had not gone down the road towards neo-liberalism. Neo-liberal solutions such as privatization, audit culture, new public management, the development of new forms of markets, etc. are conceptually connected to technocratic developments like GenAI and platform, surveillance, or intimate capitalism (Srnicek, 2016; Zuboff, 2019; Virani, 2024). As stated by Matteo Pasquinelli in his chapter on ‘Hayek and the Epistemology of Connectionism’, the connection might even be deeply structural and embedded in the very technique of GenAI. As he writes: ‘Hayek stole pattern recognition and transformed it into a neoliberal principle of market regulation’ (Pasquinelli, 2023: 183).

Though we should not come to hasty conclusions from structural and conceptual similarities, the similarities between selecting and producing outcomes via pattern recognition and neo-liberal market regulation are remarkable. Both are quantitative measures without causal, qualitative reasoning and thus open for capitalistic exploitation. The GenAI industry, with leading players like OpenAI, are heavily supported by investment capital, which is already looking for a return on investment. We might imagine something like the surveillance capitalism of Meta and Google further developed through relations to the intimate data collection of ChatGPT. Looking at how the big tech bosses team up with Trump in his second presidency in 2025 signals a dystopic future version of algorithmic capitalism, that might be more damaging than the past effects of social media. Facing this, we need alternatives such as #freedomlibertyandhiphop.

Hip-Hop AI Aesthetics

As discussed above, Kristoffer Ørum’s #freedomlibertyandhiphop works with GenAI’s tendency to redesign its application. Ørum looks back to the 1990s and the turn of the millennium to illustrate a parallel counterfactual history. However, instead of nostalgia, he uses the cultural energy of early hip-hop to create a contemporary narrative and vision. Through and beyond the themes of hip-hop and the labor movement, he investigates how GenAI renders its realism. But how to describe Ørum’s aesthetics?

One can view his aesthetics as a critical, hip-hop-infused AI realism. For #freedomlibertyandhiphop, the point is that the errors, warped faces, disfigured landscapes and architecture that do not quite match, indicate that the images are produced—and more broadly how they are produced. Through the series of images, we gain insight into how GenAI processes and portrays reality. French artist Gregory Chatonsky describes GenAI as an inductive realism that generates images from images or rather ‘images of images’. His description of how this inductive realism is derived from the images in the data model and simulates our ways of seeing, rather than representing an external reality, is a fitting characterization of AI realism:

The latter [AI] produces a new type of images and thus a new form of realism that is no longer strictly photographic, which was the dominant form since the industrial revolution, claiming a direct access, an imprint, to ontology through light, but which uses photography to make a statistical synthesis. The AI produces a strange resemblance, surreal, psychedelic, hallucinatory, it makes us enter a dream nested in a dream, the human dream of the machine observed by humans according to the loop of a self-interpretation that always differs a little more its referent. Its kitsch character consists in a circulation of clichés, biases, images already in circulation. These images have a family resemblance with visuality as a whole. They are images of images. […] When we look at an image produced by an AI, we perceive well a form of realism, but it does not refer any more to the print of an external reality, it is the synthesis of cultural and memorial data. The AI works the cultural heritage from side to side: the realism is the déjà-vu and this one, in spite of what we could believe, is not without strangeness, without surprise, without failure and hesitation (Chatonsky, 2021: n.p.).

Chatonsky’s remark about the surreal ‘strange resemblance’ applies particularly well to the places where GenAI fails most in its (photo-)realism. This includes the videos in #freedomlibertyandhiphop which demonstrate that GenAI video lacked cinematic realism in 2023. An example is the film about the shadow culture unfolding in the suburbs of Aarhus with its beautifully flickering montage, making it almost possible to follow the generation of images and the diffusion process in the changes of people, houses and landscape (see Figure 7). The video highlights how GenAI often flickers through clichés about suburban landscapes rather than resembling Aarhus suburbs in particular (Study number 58, Ørum, 2023-).

Figure 7: The voids and fissures where Aarhus meets the countryside – Kristoffer Ørum: #freedomlibertyandhiphop, study number 58 (video). Reproduced with permission of Kristoffer Ørum.

As mentioned above, we also see something similar in the portrayal of too formally dressed politicians in the background of the images featuring Danish PM Helle Thorning-Schmidt in the Danish Folketing, where she sometimes resembles an older Thatcher clone of herself (Figure 4). These renderings reveal how Stable Diffusion, like other text-to-image generators, produces simulations of our collective and cultural ways of seeing, rather than capturing an external reality. It attempts to come closer to photorealism without attaining full similarity, for example in the series of images that look more or less like Helle Thorning-Schmidt or Anker Jørgensen. Obviously, these images are the results of a longer process, where Ørum has finetuned his prompts and discarded a lot of other images.

GenAI’s so-called intelligence is fundamentally shaped by cultural bias, which becomes evident when it is provincialized and infused with hip-hop aesthetics. We do not just see, but rather are shown ways of (machine) seeing and Ørum uses this to also make us envision counter histories (Azar et al., 2021). The hip-hop universe in Ørum’s version is simultaneously colorful, full of graffiti and spray cans—some of them modified for multi-person use in line with the underlying story of collaborative culture. In more than one sense, Ørum sprays colors over a rather drab 1980s Danish reality. These colors underscore the project’s general imaginary of the labor movement’s use of hip-hop to develop the welfare state. Although it appears realistic, it is also marked as a fictional space, creating a reflective distance from historical reality. Simultaneously with the counter history, we learn to recognize GenAI imaging.

Kristoffer Ørum employs GenAI as a tool for critical, reflective realism, which happens through the way he has staged a cabinet of mirroring lenses, narratives and perspectives. #freedomlibertyandhiphop does not only portray an alternative utopian reality but also demonstrates its own mode of representation, including its provincial hip-hop strategy, and this strategy and its implementation in the general project are used to explore cultural bias and demonstrate GenAI as images-of-images. Rather than seeing photorealistic images of Denmark in the late 20^th-century, #freedomlibertyandhiphop demonstrates how this would be represented through Anglo-American image sets, developing in the process an alternative version of history.

To be precise, Ørum’s work presents a fictional, provincial Danish import and active cultural adaptation of an American culture adapted to a Danish/Nordic, Social Democratic workers movement tradition. And to add to the complexity, it shows this through the lens of Stable Diffusion’s LAION-5B Anglo-American biased image set. This mirror cabinet is used to demonstrate the production process of Stable Diffusion imagery – the GenAI machinery at oblique angles – and simultaneously to envision a counter history of the contemporary. A history that could have led somewhere other than where we are stuck today, but perhaps this history still exists as a potential? At least, as mentioned above, hip-hop has given non-white and other-ethnic cultures from the suburbs a stage that reaches far into the mainstream, where they are rarely heard, and Ørum also points towards a more tolerant Danish culture than the current ruling culture. As the Danish Jamaican Natasja wished for in her hit ‘Give me Denmark back’ [‘Giv mig Danmark tilbage’] from 2007: ‘Give me open-mindedness again’.

#freedomlibertyandhiphop points to using GenAI beyond mere imitation, automation, and replication of human qualities and perceptions. The project ultimately teaches us to see the world anew—not just through the content of the images (which primarily reflect our own ways of seeing and depicting as it is embedded in image models) but also through the ways GenAI can be used to produce alternative viewpoints, challenging its inherent Anglo-American monoculturalism. #freedomlibertyandhiphop manages to tackle GenAI from its default illusionism to a realism that includes demonstrating how GenAI processes the world and towards imagining a different reality. GenAI becomes a creative tool rather than a purely automated artist, creating images that undeniably resemble what we have seen before in our photorealistic reality, but skewed. By challenging the inherent tendencies towards monoculturalism, Ørum unlocks new possibilities for reflection and critique, transforming GenAI into a tool for alternate possibilities. Beyond humor and critical reflection on Denmark’s recent history, #freedomlibertyandhiphop hints at how we can design and think about GenAI in the future—as a creative and reflective interlocutor, rather than a clichéd and potentially harmful replacement.

Acknowledgements

Images reproduced by generous permission from Kristoffer Ørum. Thanks to the editors and anonymous peer reviewers for helpful comments and corrections. This work was supported by the Independent Research Fund Denmark, grant ID 10.46540/4256-00095B and partially supported by the Research Council of Norway through its Centers of Excellence scheme, project number 332643 (Center for Digital Narrative), and its SAMKUL project scheme, project number 335129 (Extending Digital Narrative).

Declaration on Generative AI

During the preparation of this work, Generative AI (ChatGPT, Stable Diffusion) was used for searches, experimental image generation and proofreading. After using these tools, the author has thoroughly reviewed and edited the article as needed and takes full responsibility for the publication’s content.

Competing Interests

After the creation of these artworks and the writing of earlier drafts of this article, the artist has become part of a research project led by SBP; the artist is hired as PhD at a different university and is supervised by a professor at this university.