By Sylvain Chan
Art is both fetishised and undervalued. Everyone loves watching movies and engaging with fanart from their favourite franchises. But many take for granted the labour of love that goes into the creation of this media, especially when it appears to be flawed. ‘Why is the animation in Invincible still so choppy?’, or ‘Who let the atrocious backlighting in Wicked slide?’, or ‘CGI in cinema has really fallen off since Pirates of the Caribbean’.
‘Artists are getting so lazy; doing art isn’t hard.’ Never mind how studios are sacrificing quality for profit, as layoffs, outsourcing, and generative AI-assisted production become the norm in the animation process. Never mind how fast turnover rates desensitise consumers to the arduous years it takes for art to be translated from page to screen. And god forbid, a director’s creative decision runs contrary to your own tastes.
In fact, it’s so simple to make art; you can make a short film yourself by simply typing some prompts into Sora and stringing clips together. Let’s ignore the holes in continuity, the distorted character proportions, and the absence of a cohesive narrative. This innovative technology will only improve with time, and empowers people that lack artistic talent to contribute to the conversation!
Why try — risking vulnerability, failure and wasting time — when you know you can have a machine churn something out for you in under a minute?
There is a lot to criticise about generative AI: from legal concerns surrounding programmes scraping copyrighted artwork, to the environmental damage these electricity, and water-intensive models exert, exacerbating the dire climate crisis. But on top of the ethical issue, I think we need to think twice about the cultural ramifications of normalising generative AI in the creative process and wider community.
In December of 2024, LSE and King’s College London hosted a an AI ‘Art’ Challenge for participants to consider the assistive potential generative AI has in the creative process, if engaged with responsibly and ethically. Conceptually, this is interesting, as it encourages a developing conversation around AI literacy.
However, I believe that using generative AI as a ‘collaborative’ tool — even if one were to just generate reference images or create a base to develop from — is no ethically different from appropriating generated content as your own without reimagining it. Regardless of how transformative your work becomes with the human touch, you are still utilising programmes built on stolen data, thereby improving its ability to process patterns and create ‘accurate’ output. Exacerbating the technical prowess of generative AI has devastating consequences far beyond art: from Open AI’s human labour exploitation allegations of exposing their underpaid content moderators based in Africa to traumatising material, to the ongoing deepfake pornography crisis affecting female celebrities and young girls alike. Expecting everyone empowered by this technology to approach it ‘responsibly’ ignores the unregulated, unethical mess that is intrinsic to generative AI.
Additionally, one must reconsider why artists should even resort to using generative AI in any capacity. Perhaps it does ‘save time’ and increase output, but what does this say about how we, as a society, value art? Must art be reduced to a mere statistic to meet KPI’s? While it is true many artists are dependent on illustrating for their livelihood, subscribing to the narrative that artists should be more ‘productive’ speaks more to the necessity of reform in their working conditions, rather than the fault of artists not being efficient enough.
Regarding art as a ‘labour of love’ is often at odds with the cutthroat nature of capitalism, where so many end up sacrificing their passion to tight deadlines and corporate expectations. But it’s important to remember that art is not just a skill, but an expression of humanity. Not everyone creates for money or fame.
The artistic process is one of dedication and persistence. No one picks up a pencil and instantaneously creates a masterpiece: it’s about having the humility to recognise your shortcomings and the motivation to desire improvement. As Austin Kleon strives to convey in his book ‘Steal Like an Artist’, inspiration lies at every corner. Whether compiling reference images on Pinterest, studying your favourite artist’s compositions and technique, or staring at a tree outdoors for too long figuring out how to render leaves from a distance, this dance of transforming, studying, and remixing starkly differs to how generative AI understands art: a procedure that scans and replicates patterns in a neural network, devoid of meaning.
Let’s take a look at the recent allegations Rainbow S.p.A. has acquired for using generative AI in the upcoming Winx Club reboot. For a show that “immortalises Y2k fashion”, and integrates characters’ appearances into their personality and overall story, it comes off as insulting to long-time fans of the franchise, to see the media they cherish deeply be so compromised on the quality of its character designs and promotional material. You cannot expect fans in good conscience to financially support, or even engage at any level, with work that lacks love.
I have witnessed the works of influential artists like Qing Han and Kim Jung Gi be instantly processed into generative AI models after their unfortunate passing. I have watched art students voice their frustrations and demotivation after seeing their university platform AI-generated artwork. I have heard of countless stories featuring freelancers losing clients to AI. For many at the bottom, the destructive impacts of generative AI is a lived reality. It is so frustrating to watch authorities with the power to regulate its usage, twiddle their thumbs and rhetorically hypothesise a world where creatives are supposed to adapt to a techno-fetishistic environment they never asked to endure in the first place.
However, it is also important to acknowledge how AI can prove beneficial to art. After all, ‘Into The Spiderverse’ animators documented using machine learning to predict how certain elements would look in newer frames at different angles. The Nintendo Switch 2 is expected to support upgraded graphics thanks to AI upscaling. These examples align with Sophia Smith Galer’s emphasis during a recent AI workshop for Authors: “the best AI keeps humans in the loop”, unlike specifically generative AI models which usurp the role of original creation from humanity.
Yet on average, incorporating generative AI into the artistic process does little to counter the rampant appropriation of creative work. It only reinforces the devaluation of the arts as something easily replicable and mass-produced by technology.
What, then, can artists do? Though AI has seemingly been normalised as an integral component to our society, there are still various methods of collective resistance that are worth being employed to challenge this erosion of culture. Many artists use Nightshade, a tool that ‘poisons’ their work to prevent AI models from accurately scraping its data. And rather than give ‘AI slop’ more engagement, people should turn to celebrating independent creators, such as Kiana Kahnsmith and her recent pilot animatic ‘Pretty Pretty Please I Don’t Want to be a Magical Girl’, or Studio Heartbreak’s Taglish animated film ‘The Lovers’, or even the friends around you.
The act of creating and appreciating art is an irreplaceably rewarding experience. Even if your first attempts at worldbuilding are clumsy, even if your first illustrations were crude Microsoft Paint doodles, they are infinitely more earnest, more authentic, and more meaningful to your own development than anything generative AI could comprehend.
