Generative AI – Artistic Threat or Creative Tool?
We have entered the age of AI (Artificial Intelligence) in the 2020s where we can generate artworks, music, designs and texts (Manovich and Arielli, 2024, p.10). This is changing the media, film and creative industries, reminiscent to the rise of CGI (Computer Generated Imagery). Artists are fearing the change that generative AI will bring to the creative industries, just as 2D artists feared the rise of CGI. Both demonstrate the artists’ ‘fear of the machine’ and ‘resistance to change’ (Lasseter, 1987, p.35). However, while there are similarities, it is important to note that the term ‘Computer Generated Imagery’ is a ‘misnomer’ as ‘the computer doesn’t generate the images’ (Lasseter cited in Wood, 2013, p.31). Generative AI, however, does.
Bolter discusses how AI can be used either as a tool for humans to use, or as a new medium itself (2023). This blog post examines how AI can be a tool for human creativity, rather than a medium to replace human creation. It will examine AI’s identity in medium specificity, the Gesamtkunstwerk, intertextuality, and authorship; and it will argue that AI should ultimately be a tool for creativity.
The following texts will inform the discussion: Motoarca, I.R. (2024) AI, Copyright, and Pseudo Art; Manovich, L. and Arielli, E. (2024) Artificial Aesthetics: Generative AI, Art and Visual Media; and Setabundhu, J. and Setabundhu, P. (2018) ‘Cathy Berberian in the Sea of the Uncanny: How Science Inspires Our Art’.
Medium specificity is the belief that art should be limited by medium to not distract a viewer. Lessing discussed this in his 1766 essay Laocoön, arguing that ‘painting and poetry should be like two just and friendly neighbours’ who cannot interfere with one another (2002, p.116). On the other hand, the Gesamtkunstwerk is about ‘gather[ing] up each branch of Art in its highest and most perfect expression’ (Wagner, 1849, p.17); in other words, combining artforms to encourage further immersion. Thinking about these concepts will inform if we should view generative AI as a medium itself or a tool.
The first TV series in India to be made purely with AI imagery was released 25 October 2025. Mahabharat - Ek Dharmayudh (Levanya, 2025) uses AI as a medium itself rather than as a tool used in tandem with human creation (Figure 1). While this can be considered to be a revolutionary step in AI film-making, the audience reception suggests otherwise with a 1.3/10 score on IMDB. Many users commented on the series being 'AI Slop' and 'an insult to artists' (IMDB, no date). The main criticisms are the lifeless movements of the characters. Furthermore, the series relies on retelling the Hindu mythological tale of Mahabharat which is of common knowledge to Indian viewers, therefore they focus on the spectacle of AI rather than the narrative telling of the story and coherence (which the series lacks with unrelated images seemingly being edited together without intentional thought). The results of this series suggest that AI still has a long way to go before it can stand as its own medium.
Figure 1 - Mahabharat - Ek Dharmayudh (Levanya, 2025)
If we consider that AI is a medium itself, and if we followed the concept of medium specificity, this could make AI a threat to artists as it could be used as a means of replacing roles within a pipeline to cut costs and increase efficiency. Kay argues that a computer is a medium that can simulate details of other mediums, making it a metamedium (1984). The same can be applied to AI art as it simulates other mediums through prompts and pixels. Anadol describes generating AI art as ‘[using] data as pigment’ (2020), suggesting how AI can be a medium itself. However, to consider AI art to be its own medium is ironic as it has trained on pre-existing art forms from all kinds of mediums, therefore it is not even medium specific in nature. Manovich critiques the medium specificity concept as no longer reflecting the actual functioning of culture, with AI being an example for this: “various cultural and technological developments have together rendered meaningless one of the key concepts of modern art – that of a medium” (2000, p.1).
As generative AI has an unorthodox algorithmic approach to art creating variants of pre-existing examples, which Manovich and Arielli call ‘computational mannerism’ (2024, p.10), one might argue that generative AI is a Gesamtkunstwerk. This is because it is an amalgamation of all types of art mediums that it has trained from. However, I would argue that it is not a Gesamtkunstwerk because generative AI lacks intention, a core belief of the Gesamtkunstwerk. Rather than making artistic choices to immerse a viewer, it is creating visuals based on patterns of existing styles which is ‘not creative at all’ (ibid.). Motoarca suggests that AI work should not be seen as art, but rather ‘pseudo art’ or ‘simulacra’ (2023-2024, p.500) as they are ‘a-contextual’ (ibid. p.515).
The algorithmic nature of generative AI is what Disney and Universal allege is a “bottomless pit of plagiarism” (Espiner et al, 2025). Midjourney has been caught generating intellectual properties of Disney and Universal, with characters such as Yoda (Frank Oz) from Star Wars (Irvin Kershner, 1980), Elsa (Idina Menzel) from Frozen (Chrick Buck and Jennifer Lee, 2013), and the Minions (Pierre Coffin) from Despicable Me (Pierre Coffin and Chris Renaud) to name a few (Espiner et al, 2025). In Figure 2, we can see that Midjourney has accurately captured Yoda's likeness, as well as Star Wars' environments and lighting style. If a person had created this, it would clearly be plagiarism of a copyrighted characted. Therefore, if Midjourney have stolen from big corporations to train their AI, then it brings the question of ethics into AI as a medium if it has potentially trained off images that creators have not consented to. This reinforces the idea that generative AI is not creating anything new, but it is simply creating a pixelated collage based on patterns of pre-existing images and styles.
Figure 2 - Midjourney's output of the character Yoda from the IP Star Wars (Amidi, 2025)
While we can approach generative AI as being its own medium specifically, the results will be simulations of other mediums which counteract the idea of medium specificity. Furthermore, while it is tempting to dub generative AI as a Gesamtkunstwerk, the results are uninspired and computational, lacking originality and intentionality that is fundamental to the Gesamtkunstwerk. Therefore, we can conclude that generative AI should not be considered as a medium of art, nor should it be approached as one.
J. and P. Setabundhu argued that ‘creativity always works in tandem with technology and skills’ (2018, p.4). This supports the idea that AI is best used as a tool. Anadol’s argument that AI is ‘intelligent only as long as we collaborate with it’ (2020) further supports this. Just as a pencil can be a tool for a watercolour painting, or a pen can be a tool for a pencil drawing, we can consider generative AI to be a tool in our creative processes to use as a foundational point or to use within the finished product. This approach goes against medium specificity; instead, it adheres to the Gesamtkunstwerk concept, where multiple mediums and tools are utilised for an advanced viewing experience.
As Weibel suggests, mixing media can lead to ‘extraordinary major innovations in each of the media and in art’ (2012). This is why AI should not be disregarded as a tool, as it can be used in tandem with current pipelines to enhance art. This approach is supported by Arielli’s argument that ‘technology is the development of tools extending our reach and power’ (Manovich and Arielli, 2024, p.11). Many studios are starting to experiment with the new generative AI technology to adapt and evolve with it, using it as a tool in their own pipelines. Anton Thorne (Senior CG Artist) and Katie Eyre (Senior Motion Designer) from Dyson talked about this at the 2025 Adobe Substance Days, where they explained how AI is helping speed up creative workflows for artists. Additionally, Svet Lapcheva (Innovation Director and Executive Producer) from the studio Future Deluxe argued how artists should work with what is in front of them, not against it; and that artistic creation is a dialogue between humans and systems/AI, where artists maintain control while collaborating with systems for new and creative outputs (2025). This is why AI is a tool; not a hero of a story and not a solution (ibid.).
As well as being a tool for creation, it is also being used as a tool of intertextuality. As generative art is easily accessible through Chat GPT, Midjourney and Sora AI (to name a few), it has ‘unprecedentedly intensified the pluralisation of the authorial figure’ (Rizzi, 2024, p.116). Generative AI allows an audience to be interactive with an IP. This fan-centred co-creation is a major benefit for IP holders as they generate data for their IPs (Part 2: Staying in character: Transmedia storytelling & transmedia franchises, no date). This audience interactivity is reflective of an ever-evolving process of media objects as a result of technological shifts, from creation to distribution, reception and interaction (Rizzi, 2024, p.116). A recent example of this was the AI Ghibli trend on social media, where fans could generate themselves as characters in a Studio Ghibli movie style (Figure 3). Comparing Figure 3 to Studio Ghibli's Ponyo (Miyazaki, 2008) (Figure 4), we can see that the AI image has captured the likeness of the watercolour medium in Ghibli films as well as the character design style. Serbanescu and Nack argue that AI ‘supports human creativity, opening a new era of collaborative storytelling’ (2023, p.9). Intertextuality of generative AI supports this as fans impose their own narratives on their AI Ghibli characters. Furthermore, by having AI create the images in the same likeness of Ghibli's medium and style makes fans feel like they can be in the Ghibli universe.
Figure 3 - example of the AI Ghibli trend (Schlosser, 2025)
Figure 4 - Ponyo by Studio Ghibli (Miyazaki, 2008).
However, this is not without critique; an AI generative trend such as this goes against what Hayao Miyazaki himself and his films stand for, such as environmentalism and anti-globalisation (Bhatt, 2025). As a result, the trend is an example of unconscious intertextuality as the author is not in control of how his media is being used by audience members. Kristeva explains that every media is ‘under the jurisdiction of other discourses which impose a universe on it’ (cited in Culler, 1981, p.105), further highlighting how creators cannot fully control the content of their work and how it is transformed. Miyazaki himself deems AI “an insult to life itself” (cited in Bhatt, 2025). This highlights the complexity of intertextuality, where authorship and authorial intentions are disregarded in lieu of fan interactivity.
Serbanescu and Nack argue that, since generative AI is computational, it will be without bias; therefore, collaboration with AI will mitigate emotional discomfort in projects (2023, p.6). While this should make sense in theory, it is not the case. As generative AI has trained from artworks created by humans, it has the same systemic biases and prejudices as humans. Hill critiques AI for ‘systemically [obscuring] or distort[ing] the depiction of people of colour’ (2025). She explains that AI has a default assumption of whiteness in heroic or historical roles (ibid.) (Figure 5). It is important to be aware of the flaws and biases of generative AI if we plan to use it as a tool for creation to not reinforce stereotypes of marginalised groups.
Figure 5 - Ai Generated Imagery depicting a white man with euro-centric features for a protagonist character (Hill, 2025).
Linking back to medium specificity, this is why we should not consider AI to be a stand-alone medium as it can perpetuate discriminatory biases. Therefore, generative AI is best used as a tool for human creation so it can be viewed critically and manipulated intentionally to make a Gesamtkunstwerk.
How the creative industries will change with the evolution of generative AI is still something to be observed. What will affect its implementation is if we treat AI as a medium or a tool. We should not adopt a medium specific approach with it as it simulates an array of mediums, and it has trained from all mediums; therefore medium specificity cannot inherently apply to generative AI. We also should not consider it a Gesamtkunstwerk because it is a-contextual, working without intention and context.
Instead, generative AI should be utilised as a tool to maximise human creativity and create Gesamtkunstwerks. It can also be used as a tool for intertextuality, revolutionising audience interaction with media. However, it should be used critically and with care as generative AI is not without bias and discrimination.
In conclusion, generative AI is can be described as an amalgamate zombification and Frankenstein creation of art – thoughtless, unhuman, though not fully understood yet. It is unlikely to threaten artists’ livelihoods as it lacks creative thought, and it is not inventive. However, it is likely to be used in artists’ workflows as a tool of creativity. This encourages me to experiment with AI tools in my own work to examine how effective its use-cases can be.
Amidi, A. (2025) Disney, Universal Sue AI Generator Midjourney, Calling It "A Bottomless Pit Of Plagiarism’. Available at: https://www.cartoonbrew.com/law/disney-universal-sue-ai-generator-midjourney-calling-it-a-bottomless-pit-of-plagiarism-247796.html (Accessed: 31 October 2025).
Anadol,R. (2020) ‘Art in the Age of Machine Intelligence’. TED talk. Available at: https://www.ted.com/talks/refik_anadol_art_in_the_age_of_machine_intelligence (Accessed 27 Oct 2025)
Bhatt, J. (2025) ‘Ghibli AI Trend: What Happens To Humans When We Let Artificial Intelligence Create Art?’, Medium, 29 March. Available at: https://medium.com/@Jia1303/ghibli-ai-trend-what-happens-to-humans-when-we-let-artificial-intelligence-create-art-2c672bd62769 (Accessed: 31 October 2025).
Bolter, J.D. (2023) ‘AI Generative Art as Algorithmic Remediation’, IMAGE: The Interdisciplinary Journal of Image Sciences, 37(1), pp. 195–207. Available at: https://mediarep.org/server/api/core/bitstreams/34b804d2-c446-4440-b8b8-e5b6af0370ac/content (Accessed: 27 October 2025).
Buck, C. and Lee, J. (2013) Frozen. [Film] USA: Walt Disney Animation Studios.
Coffin, P. and Renaud, C. (2010) Despicable Me. [Film] USA: Illumination Entertainment.
Culler, J. (1981) The Pursuit of Signs: Semiotics, Literature, Deconstruction. London: Routledge.
Espiner, B. et al. (2025) Artificial intelligence: Disney and Universal sue Midjourney over copyright, BBC News. Available at: https://www.bbc.co.uk/news/articles/cg5vjqdm1ypo (Accessed: 31 October 2025).
Hill, J.J. (2025) ‘Recreating Ezekiel Gillespie: Racial bias in AI art persists despite efforts to correct flawed algorithms’, Milwaukee Independent, 28 March. Available at: https://www.milwaukeeindependent.com/column/recreating-ezekiel-gillespie-racial-bias-ai-art-persists-despite-efforts-correct-flawed-algorithms/ (Accessed: 10 October 2025).
IMDb (no date) Mahabharat – Ek Dharmayudh. Available at: https://www.imdb.com/title/tt38781492/ (Accessed: 6 January 2026).
Kay, A. (1984) “Computer Software,” Scientific American, 52.
Kershner, I. (1980) Star Wars: Episode V – The Empire Strikes Back. [Film] USA: Lucasfilm Ltd.
Krauss, R. (2000) A Voyage on the North Sea: Art in the Age of the Post-Medium Condition. London: Thames & Hudson
Lapcheva, S. (2025) ‘Future Deluxe’, lecture presented at Mo:De, Ravensbourne University London, London, 27 June 2025.
Lasseter, J. (1987). Principles of Traditional Animation Applied to 3D Computer Animation. [online] Available at: http://www.cs.cmu.edu/afs/cs/academic/class/15462 f14/www/lec_slides/Lesseter.pdf [Accessed 22 October 2024].
Lessing, G.E. (2002) ‘Laocoön: An Essay on the Limits of Painting and Poetry (1766)’, in J.M. Bernstein (ed.) Classic and Romantic German Aesthetics. Cambridge: Cambridge University Press
Levanya (2025) Mahabharat – Ek Dharmayudh. [TV Series]. India: Jiostar.
Manovich, L. (2001) ‘Post‑media Aesthetics’, From White Cube to Cellspace: Image‑spaces in Modern Culture, [online PDF] Available at: https://manovich.net/content/04-projects/033-post-media-aesthetics/29_article_2001.pdf (Accessed: 27 October 2025).
Manovich, L. (2001) The Language of New Media. Cambridge, MA: MIT Press.
Manovich, L. and Arielli, E. (2024) Artificial Aesthetics: Generative AI, Art and Visual Media. [online PDF] Available at: https://manovich.net/content/04-projects/178-artificial-aesthetics/manovich_and_arielli.artificial_aesthetics.all_chapters_final.pdf (Accessed: 28 October 2025).
MidJourney (2025) MidJourney, version 6. [AI image generation software] Available at: https://www.midjourney.com/
Miyazaki, H. (2008) Ponyo. [Film] Japan: Studio Ghibli.
Motoarca, I. R. (2023–2024) ‘AI, Copyright, and Pseudo Art’, Yale Journal of Law and Technology, 26, pp. 430–526.
OpenAI (2025) ChatGPT (GPT‑5 mini). [AI chatbot] Available at: https://chat.openai.com/
OpenAI (2024) Sora, version Sora 2. [AI video‑generation software] Available at: https://sora.chatgpt.com/
Part 2: Staying in character: Transmedia storytelling & transmedia franchises (no date). Available at: https://inworld.ai/blog/staying-in-character-transmedia-storytelling-and-ai-characters (Accessed: 27 October 2025).
Rizzi, B. G. (2024) ‘The networked author: Looking at contemporary authorship through postdigital comics’, Anglica (Warszawa), 33(2), pp. 115–141. doi: 10.7311/0860-5734.33.2.07.
Schlosser, K. (2025) ‘Seattle engineer’s Ghibli-style image goes viral, and sparks some backlash over AI art’, GeekWire, 31 March. Available at: https://www.geekwire.com/2025/seattle-engineers-ghibli-style-image-goes-viral-and-sparks-some-backlash-over-ai-art/ (Accessed: 8 January 2026).
Serbanescu, A. and Nack, F. (2023) ‘Human-AI system co-creativity for building narrative worlds’, IASDR Conference Series [Preprint]. Available at: https://dl.designresearchsociety.org/iasdr/iasdr2023/fullpapers/76. (Accessed: 27 October 2025)
Setabundhu, J. and Setabundhu, P. (2018) ‘Cathy Berberian in the Sea of the Uncanny: How Science Inspires Our Art’, 2018 International Conference on Engineering, Applied Sciences, and Technology (ICEAST), Phuket, Thailand, 4–7 July 2018. Piscataway, NJ: IEEE, pp. 1–4. doi: 10.1109/ICEAST.2018.8434427.
Thorne, A. and Eyre, K. (2025) ‘Frameworks for Future-Proof Digital Twins’, lecture presented at Adobe Substance Days London, BAFTA Piccadilly, London, 9 October 2025.
Wagner, R. (1849). Art and Revolution.
Wagner, R. (1849) The Artwork of the Future.
Weibel, P. (2012) ‘The Post‑Media Condition’, Mute, 19 March. Available at: https://www.metamute.org/editorial/lab/post-media-condition (Accessed: 27 October 2025).
Wood, A. (2013). Talking About Maya. University of Kent [online]. Available from: https://kar.kent.ac.uk/37675/1/talkingaboutmaya.pdf [Accessed 22 October 2024].