Ghibli. A word that has taken the entire world by storm and sent our social media feeds into a frenzy since an Open AI software started recreating images in the typical style of the Japanese filmmaker-animator, Hayao Miyazaki. A 2016 interview of the famed co-founder of Studio Ghibli soon surfaced that generated much debate, where he had reportedly called the use of AI in art “an insult to life itself”.
These events bring two major questions to mind. One, what are the legalities surrounding AI-generated artworks, especially in the copyright domain? And secondly, what are the ethical dilemmas in the use of technology to reproduce a certain style of art?
To answer the first question, there have been numerous instances of AI-generated art over the years, since the world has embraced this infant of the digital world. Mauritshuis in The Hague had used an AI-generated image of Vermeer’s ‘Girl with the Pearl Earring’ when the original piece was loaned out for an exhibition. The Museum of Modern Art in New York too hosted an AI exhibition using generative adversarial network (GAN) that ‘taught itself’ art by browsing through the gallery archives. AI image generators like DALL-E, a word-play tribute to the esteemed painter Salvador Dali, brings ‘art’ into existence by following text prompts, using CLIP (Contrastive Language-Image Pre-training) that has datasets of images, some of which might be copyright protected. The new GPT 4o that is being used to generate Studio Ghibli-like images uses similar technology to recreate images in the distinct and unique style of the Japanese creator.
Copyright legislations have faced numerous challenges in registering artwork made by ‘non-human’ entities. One might remember the famous monkey dispute case over a selfie that was clicked by a macaque named Naruto. Although the US courts refused to grant copyright protection to non-humans like the macaque, this case holds a cue to the murky terrain in the grant of intellectual protections to art. This becomes pertinent especially when AI image generators like DALL-E are opening their platform to the market to make and sell artwork generated by AI. Since the datasets used to feed these platforms often use copyrighted images to train and basically reproduce art, the owners might be held liable for copyright violations under most national statutes. This is what happened in the case of Getty Images, who proceeded with a lawsuit against Stability AI for allegedly processing copyrighted images for training data.
Furthermore, the ‘originality’ requirement for artwork to be copyrightable would be hit when the artwork is generated solely by using AI. However, certain states allow for AI assistance in artwork like the registration of ‘suryast’ artwork, inspired from Gogh’s Starry Night, in Canada that was generated using the RAGHAV application by an Indian IP lawyer. Although India too granted copyright protection to the work, it was revoked later. In the US, the Copyright Office revoked the copyright granted to an image produced by text-to-image generator Midjourney, specifying its ‘non-human’ authorship. This shows the lack of a consensus in understanding of notion of ‘originality’ and ‘person’ in the domain of copyright law. Lack of clarity in legislations thus allow for companies to not only bypass the requirement of originality but also of consent and fair use in using such copyrighted material for commercial purposes.
Coming to the ethical question of using Artificial Intelligence to create art, music and the like, it becomes important to answer the purpose of art itself. Can AI generated art works be considered ‘real art’? Is the Ghibli trend disrespecting the legacy of Miyazaki or celebrating it? One can engage in lengthy conversations on the power of art to move humanity. It has been rightly argued by Studio Ghibli loyalists that engaging in this AI trend disrespects the sanctity, hard work and creativity of the artists involved in creating every single frame in a Ghibli movie. Reducing the Ghibli aesthetic that involved hours of painstaking labour and research to generate coloured images to a social media trend cannot be called a tribute. It is theft and an appropriation of art, since no consent was apparently sought by the owners of the AI platform from the artist to use his images for training their image generators.
The other side of the internet is debating on the need to make art accessible, a plea that is often used to counter AI generated work. It can be argued that a section of society, previously in the dark on the concept of Ghibli, having no knowledge whatsoever of the creator, has come to know of his masterpieces through the trend in question. Is this a sign of a changing time, where niche artists have the chance to showcase their work to the world through controversial trends? It can also be said that the reproduction of images using the Ghibli-inspired aesthetic is just that – an inspired artform, in no way challenging the originality of the Ghibli’s legacy but adding to it, as a fan-made tribute. This, however, without adequate protection, opens up a legal question, bypassing the consent of the art-maker and extending the limits of ‘fair-use’. The lack of legislation targeting AI image generators allows such billion-dollar companies to keep their training data opaque, thereby keeping the audience in dark on the kind of images used to train the datasets used in machine learning.
The Ghibli controversy has brought back the conversation on the need for stringent AI regulations, that would allow for integration of AI in art while respecting the demands of creators. The essence of art is in the capture, evocation and celebration of human emotions, something that our ancestors have been doing since the dawn of humanity. Miyazaki’s scathing critique of algorithmic art is based on the premise of art being reproduced without an understanding of human emotions, soullessly churning out pieces like a factory product. In outsourcing this very human quality to AI generators, we are at the brink of losing ourselves at the altar of technology. While allowing for technology to make our lives easier, let us not forget that it is art that lends us our conscience.
Anindita Bhattacharjee is a Judicial Officer under West Bengal Judicial Service. A BA.LLB. (Hons.) graduate, she has completed her LL.M. in 2023 and is currently pursuing her Ph.D. from the West Bengal National University of Juridical Sciences, Kolkata.
