Waseda WeeklyWaseda Weekly

The future of AI and humans

Since the end of 2022, generative artificial intelligence (hereinafter referred to as generative AI), including "ChatGPT", has been attracting a lot of attention. ChatGPT is an interactive AI that when you type what you want to ask into the input field, it returns an answer as if you were talking to a person. Due to its high level of convenience, it is attracting attention as a new tool that can replace traditional search engines.

How will this AI change our lives and our future? Also, how should we deal with AI? This time, Waseda University students actually experienced ChatGPT and spoke to Professor Tetsuya Ogata and Professor Daisuke Kawahara of Waseda University's AI Robot Research Institute.

*Information as of the time of interview on May 8, 2023.

What kind of tool is "ChatGPT"? A Waseda University student actually tried using it.

ChatGPT (Chat Generative Pre-trained Transformer) is an interactive generative AI released by OpenAI in the United States in November 2022. Currently, GPT-4 has been released to the public after many version upgrades, and its performance is increasing.

In response to growing interest in generative AI such as ChatGPT, Waseda University has officially announced its position on the use of generative AI.

What kind of generative AI is it that is attracting so much attention? First, we asked Shino Watanabe, a Waseda Weekly reporter and second-year School of Education, to try out ChatGPT.

Female image 01
Watanabe

Today is my first time using ChatGPT, but some of my friends and seniors around me seem to be using it. In particular, when I heard that it was being used to create entry sheets for job hunting, I was surprised and wondered, "Is it really being used that practical?"

I'm also interested in ChatGPT, so I'm looking forward to trying it out! I'll immediately ask ChatGPT the three questions I've been thinking about.

Watanabe

What is the significance of university?

Watanabe

As soon as I typed and submitted my question, ChatGPT started generating text! The text comes out as if a person was typing. It's kind of strange.

ChatGPT icon

Universities have various meanings. The main meanings are listed below.

1. Acquisition of knowledge: University is a place to learn specialized knowledge, and you can deepen your knowledge in the field that interests you. This will help you in your future career and personal growth.
2. Research and discovery: Universities are also places for the discovery and development of new knowledge and technology. Students and Faculty conduct research and contribute to the development of society as a whole by bringing about new discoveries and technological innovations.
(Original text)

Watanabe

I'm sure. I will omit it here, but there is actually a detailed explanation in the following sections.

ChatGPT icon

3.Improvement of logical thinking and problem-solving skills
4.Building relationships
5. Gaining a global perspective
6.Social contribution

(Excerpt from original text)

Watanabe

The answers are convincing, and it's amazing how they can be summarized in such a logical and easy-to-understand manner.

Let me ask you the following question.

female image 02
Watanabe

Please tell me the latest information about Waseda Weekly.

ChatGPT icon

I'm sorry, but my knowledge is up to date as of September 2021. Therefore, we are unable to provide the latest information on Waseda Weekly.
(The following is omitted. Original text)

Watanabe

that? It seems like you can't answer that. After looking into it, it seems that ChatGPT cannot currently support information after September 2021.

So, next is this question.

Watanabe

What is the most important thing in communication?

ChatGPT icon

The most important thing in communication is "respect for the other person." (Omitted) Respect-based communication includes the following elements:

1. Active listening
2.Open-minded
3. Respect the other person's feelings and position
4. Clear communication
5.Body language

(Excerpt from original text)

Watanabe

Hmm, that seems to be the case. Although I have a weakness for up-to-date information, I'm surprised that I can get so many answers so quickly.

How will generative AI like ChatGPT change the world in the future? I've been curious about a lot of things, so this time I'd like to talk to Professor Tetsuya Ogata and Professor Daisuke Kawahara of Waseda University's AI Robot Research Institute, who are experts in AI!

What will the future hold for AI and humans? Interview with two Professor from the AI Robotics Institute

Images of Tetsuya Ogata and Daisuke Kawahara

Tetsuya Ogata

TETSUYA OGATA

Waseda University AI Robot Research Institute Director/ Professor, Faculty of Science and Engineering

Completed doctoral course at Waseda University Graduate School of Science and Engineering. Ph.D. (Engineering). Since 2020, he has served as the director of Waseda University's Future Robotics Organization 's "AI Robot Research Institute." Based on cognitive developmental robotics research and deep learning technology, we conduct research on AI robots that support humans in daily life. In recent years, he has been working on research to help robots learn tasks that are easy for humans to perform, such as cooking, but are difficult for robots to perform. In 2023, he received the Minister of Education, Culture, Sports, Science and Technology's Commendation for Science and Technology Award (Research Category).

Daisuke Kawahara

DAISUKE KAWAHARA

Waseda University AI Robot Research Institute Researcher/ Professor, Faculty of Science and Engineering

Completed doctoral course at Kyoto University Graduate School of Informatics. Ph.D. (Informatics). He has been in his current position since 2020 after serving as Associate Professor at Kyoto University Graduate School of Informatics. He specializes in "natural language processing," which allows computers to understand the natural language that humans use on a daily basis. Since around 2019, he has been involved in the development of large-scale language models for Japanese. Recipients include the Minister of Education, Culture, Sports, Science and Technology's Commendation for Science and Technology Award (Research Category).

Even experts were surprised by the performance! How does “ChatGPT” work?

discussion images

At the Moonshot Project's experimental room (an area recreating a cafe) in Building 120-5 on Waseda Campus

Watanabe

Thank you, teachers, for your help today! I recently tried using ChatGPT for the first time, and was surprised at how quickly and accurately I received answers to my questions. How do you teachers view the current state and trends of generative AI like ChatGPT?

Watanabe

You'll be surprised when you touch it for the first time. We are also surprised by the accuracy of ChatGPT. However, the ChatGPT technology itself has been known among researchers for quite some time, so I expected that such a tool would eventually come out. However, it appeared about five years earlier than I expected.

Watanabe

5 years! This means that technology is progressing quickly.

Watanabe

That's right. Professor Kawahara specializes in natural language processing used in ChatGPT. What do you think of him as a teacher?

Watanabe

The AI called “GPT”, which is the basis of ChatGPT, appeared in 2018. At that time, GPT already had performance that could be said to have changed the world, and it was a hot topic among natural language processing experts. GPT became even smarter and appeared as ChatGPT in 2022. I expected the performance to improve, but ChatGPT's performance even exceeded my expectations, so I was very surprised.

riverbed images
Watanabe

It's was so. First of all, how does ChatGPT work, and what exactly is amazing about it?

Watanabe

There are many things we don't know about the latest ChatGPT algorithm because it has not been published in a paper. However, papers have been published on the version of "InstructGPT (GPT-3.5)", which is the predecessor of ChatGPT. According to it, it basically uses a method called machine learning, and this learning is done using a huge amount of text. In addition, reinforcement learning is used to create and learn question-answer pairs by hand, such as, ``If you receive a question like this, respond like this.''

Watanabe

In other words, ChatGPT is made from a combination of a huge amount of learning data and human adjustments.

Watanabe

yes. As I just mentioned, the technology itself is not novel. What's more, even though there must be significantly less learning data for Japanese than for English, for some reason they are able to reply fluently and in a way that seems appropriate. Exactly what is going on in this regard is a mystery even to us experts.

A general-purpose AI robot is being developed at Waseda University!

Watanabe

If you watch the news, you'll see a variety of developments, such as the creation of more and more services that utilize ChatGPT, and the emergence of companies introducing ChatGPT into their business operations. Are there any generative AI-related things that teachers are paying attention to?

Watanabe

I am paying attention to the fact that ChatGPT is highly accurate even in Japanese, even though it is an American-made AI. From a security perspective, there are concerns about the situation in which a single American company is in possession of information entered in Japanese, so we Japanese experts quickly investigated Japanese-made Japanese products. I think we need to develop language AI.

Watanabe

My specialty is robots, so I'm paying attention to collaborations with that field. In the first place, ChatGPT is a language model, so all it can do is "answer questions." However, recently, it has not only functioned as a language model, but can also be used to draw illustrations by combining it with other generative AI and tools.

AI that can process information in various formats such as text, images, and audio is called "multimodal AI," and ChatGPT seems to be moving in the direction of multimodality.

If that happens, I feel that the day will soon come when various data from the real world will be connected to generated AI. If I were to take in real-world data and output the processed results as robot movements, that would be exactly the type of general-purpose robot that I am aiming for.

ogata images
Watanabe

Speaking of general-purpose robots, Waseda University's AI Robot Research Institute (*1), which Professor Ogata heads, is participating in the government's "Moonshot Goals (*2)."

(*1) A research institute that works on themes such as "Solution of social problems in social implementation of AI robots" based on cognitive development robotics research and deep learning technology.
(*2) A goal set by the Cabinet Office with the aim of realizing "Human Well-being."

Watanabe

I agree. The Moonshot Goals set nine goals in the three areas of society, environment, and economy in order to solve social issues, and Waseda University is also involved. Above all, we aim to achieve Moonshot Goal 3, ``By 2050, through the co-evolution of AI and robots, we will realize robots that can learn and act on their own and coexist with people.'' A smart robot that can autonomously perform housework, nursing care, nursing, treatment, etc. "AIREC" development of We are working on this.

robot images

We are also researching the necessary movements to be able to cook (the photo shows the movements before mixing).

Image of two robots working

The Moonshot Project room in Building 120-5 is designed to resemble a house, and experiments using "AIREC" are being conducted there.

Watanabe

Currently, we are at the stage where we are trying to improve operations using hands and fingers, which are difficult movements for robots. At the same time, we are using AI to perform motion learning. In fact, Waseda University has the longest history in the field of humanoid robot research.

Watanabe

It's was so!

What are the impressions of the two Professor who saw the answers on ChatGPT?

Watanabe

This time, I asked ChatGPT three questions, including the significance of university. What do you think after looking at the answers?

Images of Ogata and Kawahara
Watanabe

Your answer is very logical and clear. It's a little too safe though (lol). But it 's actually amazing to be able to get a safe answer without fail. In the past, there have been AIs that have caused problems due to their extreme or biased answers. I think ChatGPT is adjusted so that it will give you acceptable answers.

Watanabe

However, perhaps because ChatGPT only references information up to September 2021, it did not respond to the latest information on Waseda Weekly.

Watanabe

I wish I could answer that (lol). This can be said to be a weakness of ChatGPT at the moment, but it will be addressed eventually.

Watanabe

In fact, Microsoft's search engine "Bing" uses ChatGPT technology, and it seems that it can also provide the latest information.

Watanabe

ChatGPT is also interesting because even if you ask the same question again, you will get a different answer. Maybe there is a parameter (*3) that makes the answers vary each time.

(*3) Values that affect the mechanisms of machines, systems, software, etc. from the outside.

Watanabe

In that sense, they are similar to humans. Humans also give different answers to the same question at different times.

Images of Watanabe and Ogata
Watanabe

That's right, it's the context. ChatGPT takes into account the information exchanged before the question and returns an answer that is appropriate at that time. When you take into account the context of the conversation, you suddenly start to feel more human.

Generative AI will become essential to future society

Watanabe

Teachers, what kind of impact do you think generative AI like ChatGPT will have on society?

Watanabe

This is already happening, and I think it will definitely have a huge impact on society. If you use it to support your work or studies, you can be more efficient, and you can also brainstorm ideas. In private, you can use it as someone to chat with, and there are many ways to use it. Many people are probably thinking about how to make use of generative AI right now.

On the other hand, I think you should be careful because you can do so many different things. For example, it would be better to think that information on the internet will basically be referred to by generation AI. I think some people may feel uncomfortable about that.

Watanabe

Is there a possibility of abuse?

Watanabe

Yes, there is. Cybercriminals may try to have other generative AIs create programs for computer viruses, which I believe ChatGPT is not designed to do.

Watanabe

Of course, I think companies developing generative AI are trying to avoid such risks. However, the field of AI is still in its infancy, so some people may be looking for a loophole, and it may become a game of cat and mouse.

riverbed images
Watanabe

I think the scary thing about this is that an atmosphere that generates AI should be regulated becomes excessively strong in the world. Things have improved now, but in the past there were excessive regulations that made it difficult for drone research and development to proceed as it was considered dangerous if a drone crashed.

To prevent the same thing from happening to generative AI, I think it is necessary to proactively listen to the opinions of legal experts and make appropriate regulations.

Watanabe

The benefits of generative AI are immeasurable, so it would be a big problem if it became unusable.

Watanabe

Will generative AI like ChatGPT become essential to society in the future?

Watanabe

I think it will become indispensable, just as we can no longer part with our smartphones. Generative AI will improve efficiency in all fields, and simple tasks will no longer be the work of humans. As a result, I think we will become a society in which people are more creative and focus on the things that really matter.

Use AI effectively with "critical thinking"

Watanabe

Generative AI has its advantages and disadvantages, but it will become an indispensable tool in the future. So, how should we university students face this?

Watanabe's image
Watanabe

As Professor Kawahara says, there is no other way than using generative AI such as ChatGPT. Rather, the ability to use it effectively will be needed in the future. However, there is a risk in using AI to do all the core parts of the task. Currently, there are still quite a few mistakes in the generated AI's answers, and if you submit a report full of mistakes as is, your evaluation will naturally go down.

In other words, to make full use of generative AI, you need the ability to correct mistakes in answers. To do this, make sure you understand what you are writing in your report. The image is that you can understand and write it yourself, but use AI as an aid to improve efficiency.

It is important to be aware that you are responsible for what you create. I think teachers will have different opinions on whether or not it's okay to use AI for assignments, so this is just my personal opinion.

ogata images
Watanabe

I think Professor Ogata is right. It is important to use "critical thinking" to consider whether the answers provided by the generative AI are really correct.

Watanabe

And we Faculty will also need to respond by changing the way we assign assignments and the way we evaluate them. After all, we are unable to cope with the rapid changes in generative AI.

Watanabe

By the way, do teachers also use generative AI?

Watanabe

I'm using it. I use it like a wall-based opponent, by getting them to come up with ideas and asking them what they think of the title of the paper.

Watanabe

Similarly, I have students come up with ideas and edit drafts of English papers into more sophisticated expressions.

What is the significance of universities in an era where the use of AI is commonplace?

Conversation image between Watanabe, Ogata, and Kawahara
Watanabe

As I listened to the teachers' stories, I realized that I wanted to be able to use generative AI skillfully. However, if AI becomes capable of doing more and more things, I start to wonder what the point of going to university is.

Watanabe

If you just ask a question and get an answer, it may be possible to do it with generative AI. However, at university, you can obtain cutting-edge knowledge from professors who are conducting research in various fields. I think this is a very valuable experience.

In addition, all the generative AI answers is based on information available on the internet, and only knowledge that is already known. AI cannot teach us anything that no one else in the world knows yet, nor can it discover and solve new problems. The most interesting thing about university is that new discoveries may be made through the pursuit of knowledge and research.

Building human relationships is also an important function of university. Connecting people is very meaningful, and university is a place where you can do that.

...My current answer is pretty much the same as Watanabe-san's answer to ChatGPT about the "significance of university" (lol).

Watanabe

Lastly, could you give a message to Waseda University students?

Watanabe

The advent of generative AI had a really big impact. To be honest, we cannot predict how AI will evolve or what will happen to society in the future.

However, with the advent of generative AI, it has become clearer what only universities can do and what only humans can do. People who happened to sit next to you in class, people who joined the same student club, people you did research with in the laboratory...I want you to cherish these encounters and make it a place where you can deepen your learning.

Watanabe

President Aiji Tanaka has said that `` Intellectual resilience, Flexible sensitivity, and Responsive reasoning'' are important skills for students to acquire, and I believe this is exactly right. With the advent of generative AI, even more outstanding intelligence will be needed, so I would like to see it cultivated. Please cherish the power of discovery and invention and creativity that only humans can have.

Watanabe

Personally, I had seen negative opinions about generative AI in the media, so I was hesitant to use it until today. However, listening to the teachers' stories made me want to think critically, use tools effectively, and refine my knowledge. Thank you for your time today!

Images of Watanabe, Ogata, and Kawahara

Editing/Design/Coding: Samuraito Co., Ltd.
Interview and text: Yuki Yamatai
Photography: Yuki Ide
Illustration: Yuna Toyoshima

facebookアイコン share twitterアイコン Tweet

Check the latest information on the official account!

Page Top
WASEDA University

早稲田大学オフィシャルサイト(https://www.waseda.jp/inst/weekly/)は、以下のWebブラウザでご覧いただくことを推奨いたします。

推奨環境以外でのご利用や、推奨環境であっても設定によっては、ご利用できない場合や正しく表示されない場合がございます。より快適にご利用いただくため、お使いのブラウザを最新版に更新してご覧ください。

このままご覧いただく方は、「このまま進む」ボタンをクリックし、次ページに進んでください。

Continue

対応ブラウザについて

close