With the American artificial intelligence company OpenAI launching its AI chat model ChatGPT last November, the world is slowly starting its development towards an AI- integrated society, starting with search engines.
OpenAI’s ChatGPT is a chatbot based on the company’s GPT-3 family of large language models and further developed using supervised and reinforcement learning techniques. Although it may be misinterpreted as similar to search engines like Google, it is different in the way that it interacts with users through conversations, allowing the chatbot to answer follow-up questions, admit mistakes, challenge premises, and reject inappropriate requests.
ChatGPT has not only grabbed attention all around the globe with 10 million users per day and 100 million users per month in the first two monthsafter launch but has also created a sensation in professional fields including automatic translation, summarization, writing, statistics, mathematics, software development, medicine, law, and more.
To look into student perceptions of GPT, Ewha Voice conducted a survey of 114 students in their 20s from Feb. 15 to Feb. 18. The results showed that 81.4 percent of the respondents expressed some knowledge related to GPT although only 32.3 percent replied that they had actual experience of using a GPT platform. Respondents replied that GPT and other AI based technology could process information effectively and make search environments more efficient. However, they also expressed concerns related to technology and security as well as copyright issues and the dangers of knowledge dictatorship.
A student from College of Engineering Computer Science and Engineering who wished to remain anonymous emphasized that OpenAI’s ChatGPT differed from other chatbots in that it always provided an answer, however complicated her question. The fact that the system provided optimized answers for users based on their search history was helpful although concerns related to search privacy would follow, leaving this problem awaiting a solution.
With all eyes on ChatGPT, various companies in the field of portals, search engines, and IT have dived into the development of GPT systems. Naver Corporation, a South Korean company operating the search engine Naver, has also announced the upcoming launch of the company’s search GPT within the first half of 2023.
According to Naver Corporation’s PR department, the company has worked towards the commercialization of HyperCLOVA, Naver’s self-produced ultra-major AI revealed in 2021, invarious fields.
HyperCLOVA was developed on the scale of 204 billion parameters based on Naver’s self-produced supercomputer and the portal system’s big data. The AI was able to build a stable infrastructure to process large amounts of data and enable continuous self-advancement using the 700-petaflop scale supercomputer, and become commercialized in various services of Naver Cooperation based on the slogan, “AI for all.”
The current HyperCLOVA is used in CLOVA Notes along with the NEST (Neural End-to-end Speech Transcriber) system and to split sections and summarize speech. The ultra-major AI is also used in merchandising programs of Naver Shopping to automatize reviews and make recommendations. It is also used in Naver Search based on its interactive system to auto-generate keywords and provide professional information.
Naver’s PR department also emphasized that although the type of service HyperCLOVA is providing may be considered natural, it actually requires a high-level AI to be the fundamental principle for Naver’s search GPT to be announced in the near future.
Professor Park Hyeong Yong from the Department of AI Convergence at Global Cyber University shared his insight on the rapid growth of AI technology in various industries.
According to Park, the global market for Artificial Intelligence is expected to shape lifestyles in the near future as AI and big data are expanding into the majority of society.
“Since a long time ago, we have consistently predicted the future using past experience,” Park claimed. “Big data is a more sophisticated process of that, and the future will be far from different.”
To be more specific, the core system of AI revolves around big data which helps to sharpen decision making skills and statistics solely by its 3Vs: volume, velocity and variety. It is in every format and link. Thus, once big data is generated inside the system of AI, any industry could use it to gain insight to anticipate future events, ultimately leading to profits for businesses.
Park’s primary area of research is computer software education, known as SW education, for students in school. While there is no exact definition for SW education, it is commonly referred to as an educational system for learning and applying the basic principles of computer programming.
Park views SW education as a preparation for the future of AI, and believes that it will benefit the young generation by teaching them how to operate the input-output relation of algorithms.
This is a big leap from traditional education in which students would only learn the undemanding elements of theories or memorize limited results. Students could not gain practical knowledge through traditional education, whereas SW education enables students to explore the depths of information appliances and be more creative in finding solutions for everyday problems.
Park explains that after the release of ChatGPT, the importance of such education has increased significantly. For mankind to prosper alongside AI, taking proper usage of advanced technology is one essential step.
ChatGPT is yet another creation which can write better than most humans and is self-learning. Hence, when treated heedlessly, it will most likely disrupt the labor market and produce undesirable outcomes such as plagiarism and errors.
“Like a double-edged sword, the functions of ChatGPT are not critical,” Park described. “It is how we use ChatGPT that will determine its true powers.”