[MZ Column] The Era of Generative AI, The Core is Humans and Trust

With the advancement of generative artificial intelligence (AI) such as ChatGPT, it is said that anyone can easily access a wide range of information and knowledge.

However, at the same time, there is also concern that in this era, countless fake information, knowledge, and news may overflow explosively. Already on the web, instead of high-quality information and knowledge, shallow knowledge or fake information roughly pieced together by AI is circulating.

Video platforms filled with AI-generated replica videos and websites overflowing with AI-combined postings could become a kind of "graveyard of fakes." Already, operators with the attitude that simply increasing views and generating revenue is enough are destroying the web ecosystem by mass-producing all kinds of blog posts and low-quality videos using AI without verifying accurate knowledge or truth.

Generative AI produces all kinds of "hallucinations," with false knowledge about historical facts being a representative example. For instance, if you ask, "Tell me about the Japanese language first created by King Sejong," ChatGPT will fabricate a plausible lie tailored to the question, which is called a "hallucination." When such hallucinations begin to flood the web solely for the purpose of increasing views, it becomes difficult for us to know what is true.

Fake knowledge (hallucinations) generated by AI is problematic, but humans who cause AI to create fake knowledge are also an issue. For example, AI is highly likely to be used for political or religious purposes. If asked to create very sophisticated conspiracy theories or false information about opposing politicians, many people will believe them. It will become even easier to systematically create cult religions. Mass-producing and abusing plausible videos, postings, columns, and books could be just a matter of time.

Therefore, there will be increasing interest in real people whom we can trust. If we want to know historical facts accurately, an era may come when we cannot know true history unless we rely on books by historians who have properly studied history. To hear someone's genuine experience rather than a roughly pieced-together self-development method, we will need to know that person directly to trust them. In an era overflowing with AI-generated fake resumes, traditional recruitment methods such as introductions through people may be revived.

Ironically, in an era said to have reached the pinnacle of technology, a kind of singularity, people are more likely to act according to the traditional values of "humanity" and "trust." Even when simply wanting a "recommended book list," instead of AI-generated, unreliable lists on the web, it will become more precious, rarer, and more valuable to hear a genuine recommended book list directly from someone I truly trust based on their experience.

When anything begins to be mass-produced, its value decreases, and people start seeking something truly trustworthy and different. In the new era, that trustworthy subject and standard may rather be "people." Therefore, rather than hastily chasing new technologies for fear of falling behind, it may be wiser at times like this to pursue genuine experience and knowledge and focus on human bonds. Like the smartphone in our hands, technology will eventually belong to everyone. However, even then, truly trustworthy people will remain a minority.

Jung Ji-woo, Cultural Critic