News World

Japanese startup to use AI to translate manga

Japanese startup to use AI to translate manga
Source: Unsplash

A Japanese startup said Tuesday it aims to use artificial intelligence to help translate manga comics into English five times faster and 90 percent cheaper than at present.

Mangas such as “One Piece” and “Dragon Ball” are a huge success story for Japan, with the market projected to be worth $42.2 billion by 2030, according to the startup, Orange Inc.

But it said only about two percent of Japan’s annual output of 700,000 manga volumes are released in English, “partly due to the difficult and lengthy translation process and the limited number of translators”.

But with its technology, Orange aims to produce 500 English-language mangas per month, five times more than the industry’s current capacity, and 50,000 volumes in five years. Other languages will come later.

“Compared to translation of a book, translating Japanese used in manga, which uses very short sentences of conversational language often full of slang, is extremely difficult,” said Orange’s marketing vice-president Tatsuhiro Sato.

“It is also difficult to figure out if the particular quote was actually said at a scene, or the line was a murmur inside one’s heart describing a mental landscape,” Sato told AFP.

Other challenges include the original text often being written vertically and finding equivalents for Japan’s many words to convey noises.

The firm announced that it had raised 2.92 billion yen ($19 million) in funding from major publisher Shogakukan and nine venture capital groups including the government-backed JIC Venture Growth Investments.

It said that its tool will also help the industry fight piracy in the industry, which is estimated to be worth $5.5 billion annually according to the Content Overseas Distribution Association.


About the author


Agence France-Presse (AFP) is a French international news agency headquartered in Paris, France. Founded in 1835 as Havas, it is the world's oldest news agency.

Daily Newsletter