Use of AI Art: The Online Battle Against Innovation Over Tradition

In the past few days, discussion related to the use, exposure, and proliferation of AI art online among the VTuber community has been evident, which has been used to reap social media engagement, if not potentially exploited for money.

Over the past few months, AI art generators like DALL-E, Midjourney, NovelAI, and Stable Diffusion have become viral due to its varied uses, but mostly to create weird art generations. Despite these art models having various functionalities, AI training, and source materials used–they all share the same framework: use existing art circulated online, whether they are on public domain or commercialized artwork, and use these arts to train their AI model to generate near-perfect art based on prompts given by a user.

So where did the VTuber discussions start? In the past few days, AI artwork of popular VTubers, mostly from Hololive, have been uploaded to Twitter. Some of those artworks have been posted by Twitter user @Greyfoxdelta, featuring Gawr Gura.

This has prompted Hololive members to speak out on their perspective on the use and promotion of AI art.

Takanashi Kiara, for instance, has advised her fans not to flood her art tag ‘#ArtsOfAshes’ of AI artwork, stating that “it might be hard to differentiate in certain cases and [she] would only like originally created art there.”

Kiara even had a small scuffle with a fan when it stated that artists use other’s art as reference to create new art. Kiara then questioned whether inspiration for traditional art is even on the same level as AI art.

Meanwhile, Mori Calliope and Gawr Gura have also asked fans not to flood their art tag with AI artwork, with Gura stating that it is now hard to decipher which art is made traditionally or through an AI prompt.

Furthermore, HoloTEMPUS's Axel Syrios has expressed his opposition to the premise of AI art, since it relies on other existing art work for 'training', which most of the time is not being given permission by the original artists themselves and given consent.

Other VTubers like Bao and agencies like KoMETA have already released public statements related to the use of AI artwork, stating that they shouldn’t be uploaded on their official art tags.

However, some VTubers are open to the notion of AI artwork use, albeit in more controlled situations. Hololive’s Amelia Watson, for instance, has welcomed fans to upload AI artwork under a new hashtag (the tweet has since been deleted), and dedicated a whole stream explaining the implications of AI artwork. Her openness to the technology is the polar opposite of her stance on NFTs, another industry controversially known for some NFT drops using AI artwork.

The positive sentiment also goes for Production Kawaii members Lua Asuka and Oceane Otoishi experimenting with AI art using NovelAI, a popular AI art generator specializing in anime-styled outputs. However, Otoishi clarified that using AI artwork on her end is meant as a reference for future art commissions, and will never replace traditional artwork any time soon.

Why does this story matter? The clash of traditional and modern practices of art has been going on these past few months, with questions arising on the basis of art ethics, exploitation of artwork online, and potentially giving people the power to earn through AI artwork without the training traditional artists have to go through to master their unique style.

Freelance illustrator theArtGun has explained that as AI gets better every day, it poses a threat to traditional artists, but said that there will always be people who will see the value of traditional artwork.

Meanwhile, Zenith has summed up the debacle of AI ‘users’ to those of who use Picrew, a popular Japanese avatar maker.

Issues related to AI artwork first came into mainstream attention when Jason Allen won the Colorado State Art Fair with the artwork “Théâtre D'opéra Spatial”, which was generated using Midjourney. While he only received around US$300 for the prize, his winning struck a nerve within the artist community, especially when he told The New York Times that “art is dead, dude”.

BBC asked the group Design and Artists Copyright Society (DACS) related to the rise of AI artwork, and have expressed ‘concern’ on its widespread use. Its head of policy Reema Selhi said “there are no safeguards for artists [..] to be able to identify works in databases that are being used and opt out.”

Selhi’s concern is related to the practice of some of these AI artwork generators to use existing artwork without permission from the original creators themselves. There is also concern that AI artwork generators are encouraging users to credit the generated art to themselves. For instance, NovelAI has said in its ToS that is giving its users ownership rights to those that initiated the generation itself.

However, varied copyright laws have also created mixed views on the commercial use of AI artwork. Attorney Kosuke Terauchi said that since a revision of the law in 2018, it is no longer illegal in Japan for machine learning models to scrape copyrighted content from the internet to use as training data.

As with any other trends, AI artwork seems to be impossible to be slowed down. However, they should always remember the distinction of said art styles. After all, AI art work should be able to assist, not be used to exploit.

Riva et. al. (September 2022) noted that under some circumstances, humans may conform to welcome systems that use artificial intelligence. But in most cases, subjective tasks still draws humans to seek other humans as a credible source, compared to AI frameworks. The same goes with art: no matter how complex AI framework can get to almost imitate any given artist style of any generation, the human talent will always be irreplaceable. Tangible evidence of hard work and effort will never replace any complex AI that gets in our way.

As a closing reminder, just remember what Studio Ghibli's Hayao Miyazaki said when he was presented with an AI model that generates 3D artwork: "I would never wish to incorporate this technology into my work at all. I strongly feel that this is an insult to life itself."

As much as AI-generated artwork is cool and all, it will never replace the emotional sentiment an artwork made by humans themselves. It's the reason why we have lived through so many great artistic minds over the years. It's the reason why art--real art--will always exist.

In the end, the human touch will always win over a soulless framework, no matter the intention.