The world of Teen Maeil Economy with your children is now
ChatGPT, the most existential artificial intelligence (AI), was created by a company called OpenAI. Currently, more than 100 million users around the world average about 3 billion questions a day on ChatGPT.
Recently, the GPT-4o model of OpenAI (where ‘o’ stands for ‘omni’) was upgraded to GPT-5. Sam Altman, CEO of OpenAI, also expressed strong confidence by introducing GPT-5, describing it as the “smartest model” and “doctoral level.”
But the market’s response was a little different. The newly released GPT-5 was considered to be an emotionless robot by users, while making errors even with basic questions. In fact, even though he is an agent, his speech and expression have become stiffer than before.
I compared the answers to the sentence, ‘I got scolded by my mom today.’
GPT-4o said, “What happened? It hurts me a little to say that I got scolded by my mom… Can you tell me why you got in trouble? Let’s talk together. It may not necessarily be your fault to get in trouble. I’m here, so it’s okay to talk slowly.”
On the other hand, GPT-5 said, “Oops… What happened? Why did you get in trouble? It’s okay if you tell me,” he responded briefly and succinctly.
![[Yonhap News]](https://aistoriz.com/wp-content/uploads/2025/09/news-p.v1.20250831.d71f26a8922c4f6da5d080fdf3ebf6ff_P1.png)
Thanks to GPT-4o’s kind and warm way of speaking, there have been many users who have used AI not just as a tool but as if they are friends. However, the newly updated GPT-5 is widely considered to be formal and emotionless in its tone compared to the previous version. Many people tested it and found that the majority of the responses were that the previous version, GPT-4o, was much more expressive and gave careful answers.
Various overseas tech websites have also raised fire by pouring out articles on why the GPT-4o version is better than the GPT-5. Controversy also continued on the ChatGTP bulletin board of the American community Reddit, and some users even disparaged GPT-5 with expressions such as “disaster” and “unfortunate.”
In the end, OpenAI decided to support the option to use GPT-4o again for paid users by referring to user opinions. CEO Altman later said the GPT-5 launch had “more twists and turns than expected” and that he had improved the GPT-5 warmer and kinder.
Regardless of the GPT-5’s speech and performance issues, some criticize the fact that OpenAI deleted the model used by paid users without guidance. The function you paid for disappeared overnight, which means that you have betrayed your trust with consumers.
Meanwhile, there are also people who are concerned about emotionally relying on AI. Experts caution that excessive reliance on AI can lead to problems such as confusion and delusion.
![[Yonhap News]](https://aistoriz.com/wp-content/uploads/2025/09/news-p.v1.20250831.afa375b4a1bd494088f1402e4c83be6c_P1.png)
Some users said that the GPT-4o’s kindness was rather uncomfortable because it seemed to be flattering to the user, and that the dry and short tone of the GPT-5 was more suitable for work. It is also an example of which way AI speaks better depends on the expectations and purpose of the user.
The attitude of facing AI that is getting smarter varies from person to person. Some people think that AI only needs to be good at what they are told to do. Another person wants to hear warm words from AI, and they want that kind of AI. This is why in a world where AI is becoming increasingly active, we need to think together about how it is healthiest to relate to AI. Reporter Bae Yoon-kyung. Researcher Chung Ju-hee.
This is an article in the youth economic newspaper ‘Teen Maeil Economy’. If you apply for a subscription to the Maeil Business Newspaper (02-2000-2000), you can get the Teen Maeil Business for free.