[2023 IPR Forum] "Generative AI, Securing Data Through Learning Is Key"

Lee Jungsoo, CEO of Flitto

"If we compare ChatGPT to a Hallabong (a type of citrus fruit), it is an amazing Hallabong. People will start cutting this Hallabong and selling it in various forms. One Hallabong costs 1,000 won, but by cutting it and making patbingsu (red bean shaved ice), it can be sold at a higher price. The technology to make this patbingsu is necessary, and sufficient data is required for this."

Lee Jeong-su, CEO of Flitto, said this during a lecture titled "Changes in the Translation Industry Brought by Generative AI" at the 2023 Asia Economy IPR Forum "Changes in the Capital Market Driven by Activism and AI," held on the 8th at the 19th floor of the Korea Press Center in Jung-gu, Seoul.

Lee Jung-soo, CEO of Flitto, is giving a lecture on the topic "Changes in the Translation Industry Brought by Generative AI" at the 5th Asia Economy IPR Forum held on the 8th at the Press Center in Jung-gu, Seoul. Photo by Dongju Yoon doso7@

Flitto, operated by CEO Lee, is a language data and professional translation service company. Recently, it beta-launched an online translation service called "Flitto AI Plus," which integrates generative AI technology into its AI translation engine. It differentiates itself from other services using generative AI by being based on "Flitto AI Translation," which has been trained on high-quality language data.

CEO Lee said, "Flitto possesses 300 million pieces of language data," adding, "We created AI using these 300 million language data and currently support 25 languages." He further explained, "We are enhancing the technology by inputting and training on a large amount of data," and "We are applying AI-based technology in various real-life places such as restaurants and department stores."

As such, data is the core of the AI era. However, there are copyright issues. Some may ask, "Since everything on the internet is data, why should we pay for data?" In the past, developers gathered data without considering copyright. But since 2018, copyright issues have arisen in Europe, changing the situation.

CEO Lee said, "There is a strict view on copyright, but in Korea, a different trend is unfolding," adding, "Discussions have started on lowering copyright barriers." He continued, "Globally, there is an increasing demand to relax data copyright for AI training."

Along with copyright, the inclusion of data in corporate purchasing costs is also noteworthy. CEO Lee said, "In the past, people who paid for drinking water were perceived as wealthy or foreigners, but nowadays, that perception has changed," adding, "Similarly, in the past, people did not even think about paying for data, but now it has changed." He emphasized, "American companies have now started budgeting for data-related expenses."

Labor issues related to data are also a concern. Currently, OpenAI, the developer of ChatGPT, is embroiled in a lawsuit. It came under scrutiny after it was revealed that data refinement work was conducted using African minors. In the past, children in developing countries made shoes and mobile phones, but now the era of data-related work has arrived.

He also emphasized the importance of recognizing the situation in AI translation. He said, "ChatGPT has confidently answered so far, but it is gradually trying to understand the situation through questions," adding, "By generating additional data through various questions, it accumulates more data." This underscores the importance of securing data.

This content was produced with the assistance of AI translation services.