게시물 상세

Topic

 

The Influence and Application of AI Translators on
the Publishing Industry and the Import and Export of Copyrights

 

2024.04.01

 

인공지능

 

 

AI technology is advancing rapidly and transforming the way we work. These technological developments are affecting many areas of our lives, and the publishing industry is no exception. The use of Neural Machine Translation (NMT), such as Google Translate, which emerged in 2017, as well as Naver’s Papago, DeepL, and Microsoft Translator, is growing dramatically, and the publishing industry is joining this technological trend. Its use is expanding from simple business emails to book introductions, copyright marketing materials, subtitles for book trailers, speech-to-text technology, and more. This means that a fundamental shift in working tools and business structures is taking place. This article examines the influence of AI on the publishing industry from an editor’s perspective.

 

 

에디터 윤서

 

Yoon-Seo, an 8-year editor at a publishing house near Hongdae University in Seoul, the capital of Korea, starts her day early in the morning. She dives into her research, scouring various resources to find books that will appeal to young readers. Sitting at her computer with a cup of coffee in hand, she first browses Amazon to check out the latest trends in children’s books. Among the books published in various languages, English and French titles are particularly popular among Korean children these days. While she can read and write English to some extent, she knows very little French. But, this language barrier is not a big problem, thanks to the evolution of AI translators and the availability of experts (professional translators) when needed.
Checking her email, she finds the latest rights guide and book recommendations from a local French agency. Yoon-Seo carefully reviews the list and thinks about which books would be suitable for Korean children. Before 2017, it was difficult to understand and curate foreign content, but the recent improvements in AI translators have changed the way she works.
After lunch, Yoon-Seo searches local news sites. She collects information such as new books for children and interviews with famous writers, which become important references for future import and export decisions. The language barrier sometimes hampers progress, but Yoon-Seo finds a way around it in various ways. Her English is getting better, and she has started learning French from the basics with the help of AI. She constantly seeks to present a broader world to children.
Throughout the afternoon, she organizes information and makes plans for tomorrow, then leaves the office when it is time to go home. Yoon-Seo is proud of her work in helping children experience the wider world.

 

 

Low cost, high performance

 

AI translators boast extraordinary efficiency. They are incredibly fast and cheap, even when compared to human translators. With some popular free services, you can translate mass amounts of text in a few seconds to minutes. The efficiency is even greater when it comes to the fast production of publications and the research phase of rights import and export. It became convenient to find foreign sources, get a general idea of the content, read reader reviews, and make decisions. The time spent on review has been dramatically cut down.

 

Translator Release Year Characteristics
Google Translate 2006 - Supports more than 133 languages
- Able to translate texts, images, and voices
- Improved accuracy with NMT applied in November 2016
DeepL | DeepL Pro 2017 - Supports more than 800 language pairs
- Provides high-quality, accurate translation understanding of the context based on neural network
- Provides glossary for more than 60 language pairs (DeepL Pro)
Papago 2017 - Supports 16 languages
- Able to translate texts, images, voices, and conversations
- Able to translate real-time conversations
- Developed by Naver, has strengths in translating Asian languages
Microsoft Translator 2003 - Supports 129 languages
- Able to translate texts, voices, conversations, and documents
- Provides offline translations
- Improved accuracy with NMT applied in 2017

Source: Information from each translator’s official website

 

The advantages are not limited to the fast speed and efficiency. When properly used, translators can help small and medium-sized publishers knock on doors in global markets that they would otherwise have given up on due to lack of resources. It opens the door to reviewing and producing publications in multiple languages. Small publishers used to have limited opportunities to export their works, producing once or twice a year and pitching them to agencies, but now they can organize and distribute a list of titles every month. Compared to the traditional process of translating abstracts, this significantly saves time and money, while breaking down language barriers to reach more international publishers.

 

Neural Machine Translation (NMT) opens a new horizon

 

Machine translation has a long history, starting with code-breaking in the 1950s, moving through the 2nd generation of statistical machine translation, the 3rd generation of neural network-based translation, and now the 4th generation of transformer models. However, it is only since 2017 when AI translation was introduced, that its potential to break down language barriers between countries and facilitate exchange and communication has flourished. Neural Machine Translation (NMT) is a form of translation technology based on AI and deep learning that processes the entire sentence as a unit to understand the context and provide a more natural translation. By understanding the meaning of the entire sentence instead of translating words or phrases individually, it is able to produce more accurate and fluent translations.
This is a significant improvement upon the 2nd generation of Statistical Machine Translation (SMT) methods, and the difference in performance is especially noticeable when translating long sentences or complex phrases. It has a high level of naturalness and accuracy in translating between different languages as it learns from huge amounts of linguistic data. Advances in this technology not only improve translation quality, but also speed up translation and expand the possibility of translating different language pairs. Its use is growing in the fields of global communication, multilingual content creation, international business, and education. It is also opening up new possibilities in the publishing industry, contributing to the translation of books and marketing materials into different languages and enhancing communication with international readers.

 

Generation Release Year Characteristics Strengths Weaknesses
1st generation
Rule-based machine translation
(RBMT)
Late 1950s Uses predefined
linguistic rules
Consistent translation
based on clear rules
Lack of flexibility,
poor adaptation to
new languages/expressions,
high maintenance cost
2nd generation
Statistical machine translation
(SMT)
Mid 1990s Uses massive parallel corpus-
based statistical models
Data-based flexibility,
high adaptation to
various languages
Quality depends on data,
poor understanding of context
3rd generation
Neural machine translation
(NMT)
Early 2010s Processes the entire sentence
based on deep neural network
Excellent in understanding context,
capable of natural translation
Needs massive learning data
and computing resources,
difficult to find the cause of an error
4th generation
Transformer
2017 Uses transformer
deep-learning model
for parallel processing
and understanding context
Highly accurate and natural translation,
excellent in understanding
the relationship between words
in a sentence
Needs advanced computing resources,
could be difficult to find the cause of an error
(hallucination)

 

Balanced coexistence

 

Q.
How much do you use AI translators (Google Translator, Papago, DeepL, ChatGPT, etc.) when doing market research for the import of copyrights?

 

(Based on replies from 113 editors. Survey conducted in March, 2024)

 

그래프

 

Experts in the publishing industry generally agree that the widespread use of AI translators is bringing positive changes to the market. However, due to the lack of accuracy and contextual understanding, a human expert’s review is essential when drafting copyright contracts and making important communications. To address these issues, Machine Translation Post-Editing (MTPE) is becoming more common.
The advent of AI translators is an inflection point in the publishing copyright import/export market, as they have enabled greater accessibility, reduced costs, and saved time. These benefits will shine brightly when properly balanced with professional translation.

 

 


Written by Kai Song (Editor-in-chief at Hanbit Media)

 

kbbok

Kai Song (Editor-in-chief at Hanbit Media)

#AI Translators#AI#Publishing Industry#Copyrights
If you liked this article, share it with others. 페이스북트위터블로그인쇄

Pre Megazine

TOP