[Bbanggubneun Tajagi] The Threat to Your Job Is Not AI but Your Colleagues

An Overview of ChatGPT and the Future Brought by Generative AI

Ridley Scott’s 1982 film Blade Runner depicts detective Deckard (Harrison Ford) pursuing rebellious replicants. In the film, Pris (Daryl Hannah), a replicant, utters a line that perfectly encapsulates the recent debates sparked by generative artificial intelligence (AI) like ChatGPT. Pris meets the scientist Sebastian (William Sanderson) who created her and says, "I think, therefore I am."

An image of a person working using AI created with Microsoft Bing.
Photo by www.bing.com

원본보기 아이콘

Image of a person working using AI with the AI image site Craiyon. The top image is from Bing, and the bottom one is from Craiyon, both created using the same prompt.
Photo by www.craiyon.com

원본보기 아이콘

Recently, world-renowned scholars such as Yuval Harari, author of Sapiens, have expressed concerns about the rapid pace of AI development and suggested slowing down its progress. Although this technology has the potential to bring significant changes to human history, it is advancing at a speed beyond our understanding. There are also fears that AI capable of destroying humanity, like those depicted in The Terminator or The Matrix, could emerge.

Can ChatGPT think? The Standard Korean Language Dictionary defines thinking as follows: 1. The act of considering and judging things. 2. Memory of a person or event. 3. Desire or interest in doing something. 4. The intention to do something. 5. Imagining future events. 6. Having opinions or feelings about something. 7. Showing sincerity or devotion toward a person or matter. 8. Discerning reason or logic.

ChatGPT is a large language model (LLM) based on probability and statistics. It selects and provides the words with the highest statistical likelihood. If the information is not well-learned, it is more likely to produce strange outputs. It cannot make subjective judgments; it merely reports statistics. The reason for the particularly high error rate in Korean is insufficient training data. Even with training, it can only reduce the frequency of incorrect answers but cannot become perfect. ChatGPT’s capabilities align with definitions 1 and 2. In the co-authored book The Generative AI Revolution Triggered by ChatGPT and OpenAI by Kang Jeong-su, Kim Ira, Bae Jin-beom, Seo Soo-young, Sung Young-ah, Lee Hyun-jung, and Lim Hyun-geun, it is explained: "ChatGPT is a language model that has learned to imitate even nonsense. Most human speech and writing are indefinite, ambiguous, and multifunctional. Therefore, it is nearly impossible to develop ChatGPT to distinguish persuasion, propaganda, and agitation, and to independently assess the reliability of sources to deliver only facts."

The realistic concern about generative AI like ChatGPT is related to human jobs. Some say many people will lose their jobs due to generative AI. In fact, illustrators and others have felt threatened by AI that creates images, such as Midjourney. However, as explained in this book, the threat to humans is not AI itself but colleagues who can use AI better than oneself. The ability to skillfully use generative AI will become one of the measures of competence in the future. This is because it is a revolution of tools like the steam engine, computers, and smartphones.

We will use this technology to write, search, draw, and incorporate it more into daily life. Until last year, the hottest topic in the IT industry was the metaverse (extended virtual world). However, the popularity of the metaverse has sharply declined. Could ChatGPT be a bubble like the metaverse? I don’t think so. The metaverse was a great platform but not a revolutionary tool that changed lives. ChatGPT is different. The day when I feel fear because I cannot use ChatGPT well is not far away.

원본보기 아이콘

The Generative AI Revolution｜Kang Jeong-su et al.｜The Quest｜328 pages｜19,000 KRW

© The Asia Business Daily(www.asiae.co.kr). All rights reserved.