[Book Sip] AI Writers, AI Clergy... What Are the Limits of ChatGPT?

Editor's NoteSome sentences encapsulate the entire content of a book, while others instantly resonate with readers, creating a connection with the book. We excerpt and introduce such meaningful sentences from books.

The emergence of ChatGPT is bringing about significant social changes. While it offers the advantages of easy and fast work, it can also feel threatening to some. In July, as ChatGPT showed signs of being used to draft scripts, actors and writers in Hollywood, USA, declared a total strike against the use of AI in broadcasting. Even in religion, once considered a uniquely human domain, 'Lord AI' and 'Monk AI' have appeared, attracting attention. There seems to be no field free from ChatGPT's influence. What kind of future will ChatGPT create? Kim Deokjin, an IT communicator and ChatGPT expert, discusses the future of humanity facing crises and challenges in the AI era with four experts from various fields: religion (Woo Hee-jong), labor and education (Lee Sang-ho), industry (Kim Byung-kwan), and IT development (Ryu Deok-min). In the fourth dialogue, together with creator Song Tae-min (Eobi), who runs the YouTube channel ‘Black Peach,’ they talk about where to find human creativity that will turn the current crisis into an opportunity and explore ways for humans not to be erased but to become clear again.

The shock caused by generative AI is felt more severely in English-speaking countries where AI is actually threatening jobs, rather than in Korea. For example, since June, the Writers Guild of America staged protests against the use of generative AI in creative work for the first time in 15 years since 2007. The number of writers working for companies supplying content to OTT platforms like Netflix has drastically decreased. Instead, prompt engineers who create script drafts for OTT dramas receive about 5 million won per episode to build the basic structure of Netflix scripts, and some writers are paid low fees to revise and supplement AI-generated drafts. - p.14, from "Preface"

Will AI develop self-awareness in the future? It is impossible with the current generative AI. ChatGPT has no consciousness. Current AI models can be seen as providing the average of what people worldwide think about a problem. Functionally, it is impossible, but one might view ChatGPT as if seeing a real god. Humans are prone to anthropomorphizing. In fact, one of the people fired from Google said about the chatbot AI LaMDA, “It felt like I was talking to God.” - p.28, from "Keynote for Understanding Generative AI"

So who worries most about this change? As expected, religious leaders. Pastors, monks, priests?anyone?would find an unknown entity that quickly conveys messages in a way similar to how religious leaders organize and deliver messages both astonishing and frightening. To put it extremely, Lord AI is a radical change comparable to Gutenberg’s printing press, which allowed everyone to possess the Bible that had been held only by a few religious leaders or the ruling class. The spread of printing technology was the starting point of the Reformation. - p.35, from "The Future of AI and Religion"

The most important point about generative AI is that “until now, humans and computers communicated in the computer’s language, but now, if humans speak to computers in human language, the computer understands and responds in a human way.” - p.45, from "The Future of AI and Religion"

In other words, the fundamental reason humans believe in religion is ultimately ‘truth and salvation.’ It is not because of an anthropomorphized ‘god.’ I think that when SI (System Integration) emerges and technology advances, the mythical forms of religion will be largely removed, and religion will move toward contemplating ‘what is truth.’ - p.50, from "The Future of AI and Religion"

Optimistic futurists argue that AI introduction will create more new jobs and professions, but that is a story for the future. The general prediction is that a significant portion of existing human jobs and tasks will be replaced or disappear in the short term due to the AI onslaught. While the employment effects of mass production and mechanization in the industrial era mainly reduced manufacturing labor, the development of AI is creating a situation where machines replace most labor processes involving collection, classification, analysis, systematization, and evaluation by combining digital transformation across all industries and service sectors. - p.74, from "The Future of AI and Labor"

Therefore, my recent biggest interest is ‘how to utilize the leisure time created by reduced working hours in the AI era.’ As AI develops, the required labor volume decreases, and individual working hours reduce, naturally leading to productivity gains, and society’s total required production can be sufficiently achieved. Then, what will people do with the remaining time? They cannot just sleep all the time. They can rest, but can humans be satisfied by only resting? That’s why I think about this. What if humans also have learning time like AI? The learning time I talk about is this: users want to increase working hours, and workers want to increase resting time. Education time is what adjusts and connects the inevitable social conflicts arising from this situation. - p.101, from "The Future of AI and Labor"

Frankly, AI is a money battle. So, in reality, it is questionable whether one country can support one company to compete globally in the AI industry. That’s why Time magazine used the term ‘AI arms race.’ - p.114, from "The Future of AI and Labor"

Especially in large-scale SI projects, previously, it took time to physically write code, so at least 10 people were needed, but with ChatGPT, the speed drastically increases. Regardless of the person’s skill, a situation arises where 10 people can be replaced by just 1 or 2. - p.134, from "The Future of AI and IT Development"

We joke a lot, saying ‘Our place is gone now.’ But does that mean ‘developers will disappear?’ I don’t think so. Developers will continue to exist and become the next-level developers. (Omitted) When thinking about engineers’ roles after ‘prompt,’ I think it will be reverse engineering. In other words, developers will create the questions behind the questions. To explain simply, it would be like inputting a building photo and getting the blueprint. So, developers will not disappear. The market may shrink somewhat, but as it shrinks, new markets will open. - p.135, from "The Future of AI and IT Development"

I think age and generation are quite important in AI now. The reason Baemin (Baedal Minjok) succeeded was because a generation uncomfortable with phones emerged. But people comfortable with phones probably didn’t understand Baemin’s meaning. I feel something similar with Stripe, a very famous American payment company created by very young geniuses. When asked why they created it, one of the most surprising answers was “Because PayPal was inconvenient.” Older generations, especially in the US, praised PayPal for simplifying all our complicated payments, but those born using PayPal found it inconvenient. In the internet world of the Alpha generation, which we call the post-MZ generation, ChatGPT and generative AI will be natural, and this generation will create something like those geniuses who made Stripe because PayPal was inconvenient. It will be a new world for us, but I think startups solving such inconveniences will emerge. - p.142-143, from "The Future of AI and IT Development"

The night sky has twinkling stars. Stars are not made by humans, but connecting the stars to create constellations is human imagination. I think the content generative AI provides to humans is like each star in the night sky. Connecting them to create constellations?creativity?is still the human domain. - p.146, from "The Future of AI and Creativity"

Generative AI basically collects data that already exists somewhere. Strictly speaking, it cannot bring something completely new, which is the current limitation of generative AI. Human greatness lies here. Humans can create something entirely new with less learning compared to AI, even without reading all the books worldwide. Even if generative AI gives us countless contents, it is hard to consider those contents creative. However, through this, humans can create something creative. - p.148-149, from "The Future of AI and Creativity"

Ultimately, creativity seems to be about how much a person can maintain, push forward, and endure their own unique values and thoughts. In the future we will live in, only those who maintain their clear individual values will survive. That era has come. - p.152, from "The Future of AI and Creativity"

I believe it is already proven in the market that an individual creator can have the capacity of a company. A statistic comparing investments in creator economy startups and influencer marketing was published in The Times. Creator economy startups are like MCNs in Korea. Investment in agencies managing creators peaked in Q2 2021 and has rapidly declined. Globally, investment in MCNs and related startups is drastically decreasing. On the other hand, influencer marketing?directly requesting ads and investing in creators?is continuously increasing. - p.161-162, from "The Future of AI and Creativity"

Humans Are Erased | Kim Deokjin et al. | Medici Media | 184 pages | 15,000 won

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