AI and Doujin Art: Tools, Debates and Disclosure
Image generation tools based on machine learning became widely accessible in 2022, and they reached the doujin and fan art world almost immediately. Few topics have divided the community as sharply in recent years. This article describes how AI tools are being used, how stores and events have responded, the arguments made on different sides, and the growing consensus that disclosure is essential. It aims to present the debate fairly rather than to settle it.
What the tools do
AI image generators produce images from text prompts, reference images or both. Models trained on large collections of images learn to reproduce styles, compositions and character features. Modern tools can:
- Generate finished-looking illustrations from a short description.
- Produce variations on a character design.
- Apply a style to a rough sketch.
- Fill in or modify parts of an existing image.
- Upscale or clean up artwork.
Anime and manga styles have been among the most popular uses, partly because of the size of the fan art community and the amount of existing art available online.
How AI is used in doujin and fan art
Use varies widely, from fully generated works to small assists within a traditional workflow.
Fully generated works
Some creators release illustration collections or comic-style works made primarily with AI generation, sometimes with manual editing. These have appeared in large numbers on digital stores.
AI-assisted workflows
Other artists use AI as one tool among many:
- Reference and ideation – generating rough pose, lighting or costume references before drawing by hand.
- Backgrounds – producing or adjusting background images that are then traced or painted over.
- Colour exploration – testing palettes before final colouring.
- Clean-up and upscaling – processing hand-drawn artwork.
The line between “AI work” and “AI-assisted work” is not always clear, which is one reason disclosure rules have become important. The article on making your own doujinshi discusses where such tools might fit in a small project.
Responses from stores
Digital stores were the first to face a flood of AI-generated submissions. Their responses have included:
- Labelling. DLsite, among others, introduced labels or categories identifying works created with AI, so readers can see this before purchase and filter accordingly.
- Separate sections. Some platforms list AI-generated works separately from hand-drawn works.
- Submission rules. Platforms have set rules on what may be submitted, including restrictions on imitating identifiable artists or using copyrighted characters in particular ways.
- Content enforcement. The same content rules apply to AI works as to drawn works, including the requirement that all characters are adults and that local law is respected.
Policies have changed several times and continue to evolve, so creators and readers should check each store’s current rules.
Responses from events
Doujin events have also had to decide how to treat AI-generated works. Approaches vary between organisers and over time, and may include requiring disclosure, restricting AI-generated works to certain areas, or excluding them. Events that focus on original hand-drawn work have tended to be more cautious. The Comiket and doujin events article covers event rules more generally; circles should read the specific rules for each event they apply to.
The debate
Concerns raised by artists and readers
- Training data and consent. Many models were trained on artwork collected without the artists’ permission, and many artists object to their work being used this way.
- Style imitation. Tools can imitate a specific artist’s recognisable style, which many regard as unfair even where it is not clearly illegal.
- Market saturation. Large volumes of quickly produced works can bury hand-drawn releases in store listings.
- Income. Artists worry about pressure on commission prices and sales.
- Quality and consistency. AI images can show anatomical errors and inconsistent details across pages, which matters in sequential storytelling.
- Content safety. Generators can be misused to produce prohibited content, which places extra responsibility on platforms.
Arguments made in favour of use
- Accessibility. People with limited drawing ability, time or physical capacity can realise ideas.
- Efficiency. Artists can speed up repetitive tasks such as backgrounds or references.
- Experimentation. Tools can help with exploring compositions and colour quickly.
- Tradition of tools. Supporters point out that manga has always adopted new tools, from screentone sheets to digital software; see manga art techniques.
Legal uncertainty
The legal position of AI training and AI-generated images differs between countries and remains unsettled in many areas, including copyright in the output and the lawfulness of training on copyrighted works. This article does not attempt to summarise national laws, which are changing.
Why disclosure matters
Across the debate, there is broad agreement on one point: readers should know what they are buying. Clear disclosure:
- Lets readers choose according to their own preferences.
- Protects creators of hand-drawn work from unfair comparison.
- Keeps platforms and events able to apply their rules.
- Builds trust between creators and audiences.
Good practice for creators includes stating plainly whether a work is fully AI-generated, AI-assisted or entirely hand-drawn, and describing the extent of manual work.
A note on this site
Disclosure: avhentai.net is run by the team behind AiNudeGirls, an AI image service. Readers interested in what current tools produce in anime styles can browse the AI anime gallery on AiNudeGirls, which is entirely AI-generated and shows fictional adult characters. This wiki itself does not host images and aims to describe the debate on its merits.
Summary
- AI image tools are used both for fully generated works and as assists in traditional workflows.
- Stores such as DLsite have introduced labels and rules for AI-generated works; events vary.
- Concerns centre on training consent, style imitation, saturation and income.
- Supporters cite accessibility, efficiency and the long history of new tools in manga.
- Clear disclosure is widely regarded as the minimum standard.