Korea’s webtoon industry sets sight on generative AI for dramatic growth

2023.03.10 10:47:02 | 2023.03.10 10:52:28

Webtoon images created through the Onoma’s AI platform Tootoon engine [Photo provided by Onoma AI]이미지 확대

Webtoon images created through the Onoma’s AI platform Tootoon engine [Photo provided by Onoma AI]



The webtoon industry is taking a keen interest in the generative artificial intelligence technology as it could attract more creators to webtoon platforms and bring dramatic growth to the market should it become more accessible to potential artists.

Domestic startups in the Korean webtoon market, world’s biggest, have been making various attempts to use AI and are expected to expedite the birth of AI artists.

Contrary to the existing AI technology that learns data and patterns to understand the target, generative AI creates new outputs through comparative learning from existing data. Image generation, a subfield of generative AI, can generate and color unlimited images on demand. This means you can be a professional webtoon artist without a special talent for drawing only if you have intriguing materials and stories. In other words, you can create a story using ChatGPT and make it into a webtoon using an AI image generator.

According to the information technology industry on Thursday, Onoma AI, a Korean generative AI startup, will release a piece jointly created by a human webtoon writer and AI on major platforms as early as the end of this month. AI has played a significant part in creating the webtoon based on a human writer’s storyline, such as storyboard jobs, sketching and coloring, which were performed on Onoma’s AI platform Tootoon.

“We will usher in a way that allows anyone to become a creator using AI technology,” said Song Min, chief executive officer and founder of Onoma. “We plan to shortly develop and release a technology that creates texts using ChatGPT and converts them into short-form videos through the Tootoon engine.”

Tootoon creates webtoon characters, clothes, background images and storyboards according to written commands and will be open to the public free of charge in the middle of this month. If a user, for example, types in “several students leaving a classroom,” a cartoon image that matches the description will be generated and the webtoon can be completed once lines are added.

Kakao’s image creation model Karlo [Images source: B^ DISCOVER]이미지 확대

Kakao’s image creation model Karlo [Images source: B^ DISCOVER]



Making descriptions more specific will help AI generate more detailed images. AI also instantly creates webtoon backgrounds, which takes hours for writers to complete. There are no copyright infringement issues as all images created this way are intellectual property rights of Onoma, the company said.

Onoma AI has earlier released Iconoci platform for free, which creates images using an AI model that has learned 400 million images and trained with 1 million icon data.

Webtoon and animation industries are quickly adapting to the latest change. Famous cartoonist Lee Hyun-se has started developing his own AI technology, first of its kind for an active cartoonist. He will make AI learn 4,174 cartoons he has created for 44 years. The industry expects to see his new creation generated by AI as early as in the second half of this year.

The Electronics and Telecommunications Research Institute (ETRI) is also developing Deeptoon, an automated webtoon creation technology, with the goal for a release in 2024. A human artist will be engaged in plot development and sketching, and AI will handle the rest.

Naver Corp. and Kakao Corp. are also accelerating their moves to join the trend. Naver Webtoon Corp. introduced the Webtoon AI Painter system in 2021 and has set up a Webtoon AI division to respond to the changes.

Kakao plans to create an AI startup ecosystem by creating a fund worth at least 10 billion won ($7.6 million). The company will actively share Kakao’s image creation model Karlo and has distributed B^ DISCOVER, Korea’s first AI image generator application, free of charge since October last year. It also released on Tuesday the open beta version of B^ EDIT, a web service using Karlo to translate imagination into an image.

Still, copyright infringement issues related to data used to train AI remain unresolved. Visual media company Getty Images last month sued AI company Stability AI Inc. for using its photos without permission or compensation.

By Hwang Soon-min, Kim Dae-gi, and Choi Jieun

[ⓒ Pulse by Maeil Business Newspaper & mk.co.kr, All rights reserved]