XIE Zhihai “How to be together with ChatGPT?”

ChatGPT (generative AI) became a topic in Japan this year. While we could not fully understand in Japan, it is said that in the U.S. where GPT was developed, their local government decided on its access control in schools. Some teachers have classes named “How to Use ChatGPT” to encourage using it proactively. In the academic world, ChatGPT is a threat. I am not so strong at the new functions of smartphones and home electrical appliances. But I am not in the position to ask students to teach me what ChatGPT is. Or rather, I must reconsider myself about school lessons, problems and countermeasures against tests. It is said that ChatGPT is a shock not only to academic fields, but to all ‘white colored peoples as well (Weekly magazine “Toyo Keizai”). AI can be accepted by not only professionals in the technology field and AI researchers but by everybody. In the year 2023, AI became familiar through ChatGPT. I think ChatGPT declared “You cannot turn back now”.

 

As we cannot be afraid of saying threatened or shocked, I tried to create my account. I registered my email address, name, date of birth, and cell phone number. And I put in my authorization code, which was sent to my cell phone. It means people who have cell phones can create an account. It was not so difficult, and I could open my account easily and use it now. It would be timesaving, if my ChatGPT could complete the essay undermentioned. But, as it was not so easy, I input my question, “I like to teach my university students important notes when they use ChatGPT.” Then, I got a brilliant reply in twenty seconds:

 

Please note following when university students use ChatGPT:

 

Confirmation of sources:

As ChatGPT has no information before the year 2021, there is a possibility that we cannot give up-to-date information. As to important information, we recommend you confirm it from in reliable sources.

 

Protection of personal information:

ChatGPT is not a tool to protect secret information. Please do not share personal or private information.  

 

Perception of bias:

As ChatGPT learns from training data, it is possible to be influenced by bias. When you judge information, it is important to keep your objective viewpoint and make your original research.  

 

Time management:

ChatGPT is very useful tool. But it is important to manage time properly. Do not depend on others too much and cultivate your own ability to solve problems yourself.

 

Critical thinking:

Answers by ChatGPT are made by machines and are not always correct. You are requested to have a doubt and try to have critical thinking.

 

The above are important points about which university students should take care when they use ChatGPT. Keeping those points in mind, please take advantage of this tool.

 

I was surprised at how smooth and beautiful Japanese is! Those sentences look like the ones that human beings (of course, Japanese) studied and wrote hard, and they are different from the ones that AI made awkwardly. I was also surprised that those sentences were itemized so that everybody could understands them easily. I did not ask ‘itemized’. The most surprising point is the fifth advice, rather than the cautionary point. We are requested to have ‘critical thinking’, which we, as teachers, tell students to have in the classes. We are scared that ChatGPT will see through teachers’ techniques.    

 

I got punched when I heard the opinion of Mr. Jensen Huang, CEO of NVIDIA, a major semiconductor maker in the U.S. At the keynote speech of the development conference, he said “Generative AI started ‘the iPhone moment’ of AI”. I am ashamed of having ignored his opinion first. But I can understand now. It is terrible to use AI too much on smartphone. But it is more terrible that you do not use it at all. It is generative AI, starting with ChatGPT.

 

The CEO of ‘OpenAI’, the original developer of ChatGPT, strongly urged regulation of AI at public hearing of the U.S. Federal Parliament. I do not know its meaning. As he recognizes its dangers and disadvantages most, the Government has started discussions with representatives of generative AI developers, and they are in the stage of developing concrete guidelines about the use of AI. Actually, Microsoft executives who offered ‘Bing’ (a search engine) said “We welcome the function of the Government and authorities concerned who offer AI safely to people”. It seems the Government and enterprises that develop AI are moving forward. How about in Japan? Weekly magazine Toyo Keizai summarizes their opinions in their special issue “Revolution of Work Hacks by ChatGPT” about the use policy in big companies saying that “Japanese companies are very cautious”. Big companies starting with Panasonic say, “We can use ChatGPT on the condition of our internal rules”. It is a good point about Japanese companies that they make internal rules and employees obey the rules.

 

How about in universities? The university where I am working follows the guidelines of the Ministry of Education. However, at the actual site of education, we cannot wait for it and have begun to consider generative AI. When I searched ‘Waseda University’ or ‘ChatGPT’ by Google, I could hit a page ‘About the Use of Generative AI’ on the university’s home page. It is a message to students and teachers. The message is easily understandable and explains how to use ChatGPT. It explains the strong and weak points of Generative AI, how to use it, and the reasons why we cannot use it.

 

I do not understand yet whether everybody can utilize Generative AI, which enterprises and schools can access easily, and utilize it on the condition that everybody can access it. When we began to use Twitter or LINE at first, everybody jumped on them, and then, problems appeared depending on how to use them. In the case of ChatGPT, a representative of the developing company himself appealed in public for regulations on how to use it. It means there is a possibility that “it may bring risks into society and human beings”.

 

When I used ChatGPT, I realized I should be cleverer than ChatGPT, which can manipulate words well like human beings. I think strongly that we should not rely on ChatGPT too much and should not be governed by it. At this moment, the ideal way for users who should be wiser than ChatGPT is mentioned in Toyo Keizai as mentioned above. “Let’s fear, know and use properly”. Generative AI is always being developed somewhere or everywhere in the world.

 

 

SGRA Kawaraban 742 in Japanese (Original)

 

 

XIE Zhihai/Professor of Kyoai Gakuen University.  

 

 

Translated by Kazuo Kawamura

English checked by Sabina Koirala