The World Has Taken a New Turn with the Advent of Chat GPT1)
Artificial intelligence (AI) is gradually developing rapidly beyond the limits of technology. Chat GPT (Generated Pre-trained Transformer) is attracting great attention around the world with its breakthrough performance. This is an AI chatbot specializing in conversation based on "GPT-3.5" language technology. It was released on November 30, 2022, by Open AI, an American AI research laboratory, and it surpassed 1 million users a day in the first five days of its launch. Open AI also released "Chat GPT Plus," a paid version of Chat GPT, on February 1, 2023. Chat GPT can not only answer questions simply or provide useful information, but also give creative ideas or even suggest solutions to problems. Therefore, it can perform tasks in a wide range of fields such as writing papers, translating, and coding. MIT Ph.D. students found in their research that people such as data analysts and HR professionals could reduce the time needed to write press releases, short reports, and emails by 37% or 10 minutes through chat GPT and the results of the task were also superior. Due to the craze of Chat GPT, domestic companies have begun preparing services based on supergiant AI. Naver announced that it will unveil the super-large language model HyperCLOVA X as a Korean version of Chat GPT in July. Kakao will soon release an AI language model that can understand Korean through a dictionary and context based on its connection with Kakao Talk. As such, Chat GPT is promoting development of the AI market.
It has high performance and utilization, but it is difficult to ignore the ethical issues. The ethical risks of AI have been consistently pointed out so far, but the scope of its side effects is expanding. First of all, problems related to personal information protection and data security, such as copyright of learning data, personal information leakage and plagiarism, are being discussed. In addition, if the wrong information created by Chat GPT is shared on the Internet, this wrong information can be absorbed back into the model and data contamination can worsen. This can lead to the risk of producing biased and unethical answers. Chat GPT was originally planned to be developed secretly, internally verified for safety, and introduced as a complete service. However, the developers felt limited in continuing development only within the company, so they decided that it was a more realistic way to open Chat GPT to the public and institute a verification process. Greg Brockman, the chairman of Open AI, said that "If humans manage, supervise, and provide high-quality feedback, more reliable AI can be created." He believes that continuous research and monitoring by humans is needed to compensate for the side effects that generative AI can cause. The size of the generative AI market is expected to grow from 13 trillion won in 2022 to 142 trillion won in 2030. An important key to responding to the changes that tremendous growth of AI will bring is constant reviews and systematization of the system.
1) Choi Won-jin, "Chairman of Open AI "Humans Must Manage, Supervise, and Validate Chat GPT"", NEWSPIM, April 19, 2023