#graphic #manga #comic #fan #Japan #softpower #AI #Midjourney #DALLE #ChatGPT #OpenAI #art #prompt #largelanguagemodel
In the current scenario, where OpenAI ChatGPT has taken the digital world by storm, artificial intelligence can be observed penetrating all walks of human life. No one can disregard the hues and cries, as well as the euphoria that has been engendered in its wake. In this context, an artist who goes by 'Rootport' has created a manga titled Cyberpunk Peach John, which hit the shelves in March 2023. The artist has used the AI tool called 'Midjourney' to create an entire manga text of 100 pages, which took him about six weeks! Naturally, fandom has borne a divided opinion on the content and quality of Cyberpunk Peach John; much like the rest of humanity, in regards to the pros and cons of such a technology.
To those unaware, images on AI platforms such as Midjourney, DALLE and others, are created with precise and specific 'prompts'. For instance, to generate the image of the male lead, Rootport had apparently used the text-to-image prompt 'pink hair, Asian boy, cyberpunk, stadium jacket, manga'. To cite another example, here I have an AI generated image, which I created on the tool called PicSo, by using the prompt 'anime girl, teenager, light brown hair, red hair ornament, long yukata with sunflower print, festival background, paper lantern in hand, praying at a Shinto shrine, fireworks in the sky'.
Not perfect, but pretty neat, right?
Manga/art/creative industry has had mixed opinion about such AI generated creations. Understandably, the major concern has been about the 'authenticity/human touch/artistic uniqueness'. Author Jonathan Clements in his article (https://blog.alltheanime.com/manga-cyberpunk-peach-john/) has denounced Cyberpunk Peach John as a hack-ish text with inconsistent art. Perhaps, he is not entirely wrong as well.
However, one must remember, manga/anime/allied merchandise is after all, an industry. To those familiar with the theories of Adorno and Benjamin, are aware of the ideas such as 'culture industry' and loss of 'aura' for 'art in the age of mechanical reproduction'. When mass produced, artistic/creative content cannot avoid certain repetitions. Even without AI, several manga artists use stock character sketches and visual details, repeated tropes, and similar motifs across their oeuvre. I have encountered heroes that look exactly alike, in different manga texts, created by the same mangaka. Creating a pool of images is a common practice that dates back to the era of Osamu Tezuka. With the advent of different artistic softwares, several mangakas have resorted to digital sketching methods. Rootport used a similar technique of creating an image bank, aided by AI.
Japanese manga artists work under immense pressure to meet deadlines, which reoccur every week/fortnight/month, as per the serialization schedule of a particular text. (In this regard, the anime series titled Sekaichi Hatsukoi furnishes a detailed glimpse into the realm of manga publishing, wherein it is aptly observed that the nexus the artist- editor- printer can never catch a break.) There are actual records of editors locking manga artists up in hotel rooms with all necessities provided, disallowing them to come out until the series instalment is handed over! Considering the corporate work culture of Japan where 'karoshi' (death by overwork) is not uncommon, one perhaps can readily believe such instances. Herein, AI might come in handy, rescuing the artists from such a haranguing circumstance. Based upon a original sample sketch provided by the artist, AI can generate a pool of images, involving the same figure/character. DALLE actually helps create variations out of a single image that is uploaded by the user. (https://platform.openai.com/docs/api-reference/images/create-variation?lang=python)
One of the major challenges of the cybernetic nexus has been plagiarism. Unethical uses of manga art, piracy, violating copyright issues have been tailing the industry and the fandom. With constructive use of AI, one can generate their own visual manga content, erasing the need to misappropriate others' sources. Or perhaps, if one has a manga plot/storyline in mind, but is not comfortable to draw or collaborate with another artist, AI can come to rescue.
Like every development in technology which was designed to make human tasks easier, AI is the next step in the journey. If used safely, AI has the possibility of redefining manga artistry, lessening the burden on the professional artist who has the constant challenge to be commercially productive, without compromising on quality. That is not such a devastating premise, is it? Looking forward to more sophisticated AI-generated manga!
Stay tuned for more @Dr Otaku :)
Very interesting read! I think this makes the entire process easier for mangakas who aren't that good at drawing like the creator of One Punch man. One should also note that the ideas for the character design like facial features can be given to the AI while designing the character. So while it is create by a non human entity, the idea of the human is still taking some form. So long as the world building, characterisation and the dialogues are made by the human, i think AI can be used successfully without diminishing the artistic value of the work. A question to ponder about would be that if this practice continues, and with the consequent advent of better AI…