Table of Contents
- Contact points with AI and philosophy
- Rethinking “what is reality and what is the world” through AI research
- AI expands philosophy
- Differences between AI and Philosophy | “What is the mind?” “What is intelligence?”
- Intersection of AI research and philosophy
- Mutual Development of AI and Philosophy |
- Mutual understanding of each other’s fields is necessary
- Now is the time to learn philosophy
- in conclusion
Contact points with AI and philosophy
Rethinking “what is reality and what is the world” through AI research
ーーMr. Taniguchi, you didn’t have much of an image of philosophy, but do you have any connections?
Mr. TaniguchiIn the first place, I recognize that philosophy is basically something that everyone does.
My field of specialization, symbol emergent robotics*, is a constructive research based on the concept of symbol emergent systems. So I consider myself to be very involved in philosophy.
*Research to structurally understand the linguistic communication that humans acquire as they grow by having robots perform it.
Mr. HiraiThat kind of thinking is already a philosopher (laughs).
In AI research, we often have to reconsider the obvious , such as “What is reality and the world?”
Mr. TaniguchiThat’s what makes it interesting.
As I research AI, there are times when I think, “Humans are amazing.” It’s very interesting to think that even if I put all my effort into building and implementing a model, people end up doing these things unconsciously.
AI expands philosophy
At the beginning of this interview, Mr. Taniguchi said, ” I believe that AI expands philosophy, ” and talked about how he perceives his own philosophy.
ーーWhat do you mean by “AI expands philosophy”?
Mr. TaniguchiPhilosophy is speculatively discussed in various ways, but sometimes the content itself becomes fluffy.
By creating a computational model, that is, a model that can be operated using a computer, we can describe in a different form what kind of phenomenon is being discussed. For me, instead of pen and paper, I use computers and robots to draw and study human intelligence.
ーーBy using AI, the form of philosophy output is changing. Mr. Hirai specializes in philosophy, and I would like to ask Mr. Taniguchi what he thinks.
Just like Mr. Taniguchi, I believe that philosophy applies to many things related to AI.
Since 2016, I have published three collections of articles, including ” Dissecting Bergson’s Matter and Memory .” Bergson was a philosopher who studied consciousness and the mind from the perspective of time. 100 years ago, he had no choice but to think about what consciousness and mind are. Bergson emphasized the developmental perspective, but in fact, his idea of “creating (consciousness and mind)” was not viable in the first place.
In modern times, the number of options has increased, and it has become possible to actually create things using a constructive approach (a method that aims to understand by creating), and other approaches such as collecting data from living brains in real time. It is now possible to combine Until now, there was only one approach to the problem, but now there are two, three, and three-dimensional responses to the problem of consciousness. I think that this is a change in the human consciousness problem that has occurred in the last few decades, and it is one of my major pillars.
Differences between AI and Philosophy | “What is the mind?” “What is intelligence?”
ーーI feel that both AI and philosophy are thinking deeply about what the world takes for granted.
Mr. TaniguchiI agree. On the contrary, if I try to put it in a more challenging way, I would ask, ”
What is the difference between AI research and philosophy? ” If you think about it with an open mind, at least part of the act of philosophy is to think about life, your mind, your consciousness. And AI is the act of thinking about what intelligence is. Ultimately, philosophy is to think about “what is the mind?” and AI is about “what is intelligence?” When I think about it that way, I actually
feel that the fundamental question is the same .
ーーWhat do you think about the difference between AI and philosophy, Mr. Hirai?
Mr. HiraiPhilosophy is a vast field. I am researching philosophy related to consciousness and the mind, but there are many fields such as what time, existence, and language are, and discussions have accumulated in each field.
Within the theme of ” artificial intelligence ,” there are many points of contention that overlap with various fields of philosophy, such as physicality, cognition, and what intelligence and creativity are.
That is the interesting part of the chemical reaction (between AI and philosophy).
Especially in the discussion in engineering, I think that there are many cases where the discussion progresses in the direction of how to shape it by accumulating base technology and knowledge. But on the other hand, we already have a conventional philosophy. Since we have been working in a top-down manner from various conceptual visions, I think that by successfully fusing the two, we may be able to see perspectives that were previously invisible.
Moreover, philosophy is about riddles. There is a distinction between “problem” and “mystery”. (Cola Diamond (Philosopher, 1937-)). To put it simply, the problem is a math test question. There is a formula and you use it to get the answer.
On the other hand, challenging the question of consciousness and mind is not a problem, but “raise a mystery”. There is no solution (formula) to the mystery yet. So it doesn’t make much sense to apply general equations and formulas. A mystery only makes sense when you know how to solve it. So is the mystery of the Sphinx. The interesting thing about riddles is that even though you don’t know how to solve them, you know that it’s a mystery that you should try.
In that sense, I think that the people who are researching the problem of ” artificial intelligence ” at the forefront of today are also facing the same “mystery and mystery of the problem itself . ” I think that’s where it gets interesting.
ーーI would like to hear your thoughts, Mr. Hirai.
Mr. TaniguchiI think so.
When I took the podium and had discussions at events such as the Society for Applied Philosophy, I was reminded again that philosophers are good at asking questions, and they keep asking questions one after another. As if to try not to bring the discussion to an end (laughs).
In engineering, we have to solve “questions”. Value is created by solving. Therefore, I think that engineering researchers tend to feel hesitant about asking philosophers one after another unsolvable questions. However, I believe that the philosophy of facing the “mystery” that Mr. Hirai mentioned has an important meaning.
Intersection of AI research and philosophy
Looking back historically on the relationship between AI research and philosophy, it has generated a lot of controversy. For example, discussions such as the “frame problem” advocated by John McCarthy and Patrick J. Hayes and the “symbol grounding problem” advocated by philosopher Stevan Harnad are being revived in the present age of the third AI boom.
So, where and how are AI research and philosophy actually being discussed? We also asked about the relationship between AI and philosophy.
Mutual Development of AI and Philosophy |
ーーWill philosophical research accelerate further with the advent of AI and future developments?
Mr. TaniguchiNaturally, it will be accelerated by AI research, and I think it must be accelerated. From a global perspective, Japan’s efforts in AI research are relatively weak. I think there are various reasons for this, but I think one of them is
that the discussions that include points of contact with philosophy are not very strong .
Japan has technology, but my impression is that there are few people who can pursue philosophical discussions. Looking not only at AI research, but also in Japan as a whole, I believe that without researchers who are well-balanced between the humanities and science and technology, the foundation for academic discussion cannot be built. That’s why there are no impactful papers about the idea itself in AI. I can’t write in English.
Even now, much of the budget for innovation strategies is allocated to science and engineering. Furthermore, I think that the current situation is that it is difficult to develop philosophical fields because they seek practicality in a short period of time.
ーーAre there opportunities for interaction between AI researchers and philosophers?
Mr. HiraiYes, but I feel like there should be more. I think that there should be more places where people in different positions can talk.
The reason why it is difficult to find time for exchanges may be that I am busy with daily research, writing papers, and teaching. Perhaps there are many AI researchers and other science-oriented people who have such philosophical motivation and interest.
Mr. TaniguchiThere is a gap, and I think there are many dialogues where people stand on both sides of the gap and talk to each other. Such dialogue tends to assume that there are already differences between the two and that they are not connected. Talk and leave, talk and leave repeatedly.
Instead, in the case of AI and philosophy, I think it is necessary to forcibly step into each other’s territory (overcoming the gap) and discuss with a stance of trying to do something about it. AI researchers must also understand that.
In addition, I think that the reasons why there are not many opportunities for dialogue are extremely diverse and intertwined.
When asked what we do research for, the word ” practicality ” comes up. I feel that the problem is that “technological development for economic development” is accepted uncritically as justice.
Economic development is of course necessary, but I feel that the philosophical perspective of “understanding the world and the mind” is considered inferior when considering what is important as civilization.
I believe that this difference in perception is one of the reasons why it is difficult to create opportunities for dialogue between both parties.
ーーIt is true that things that are immediately usable and highly practical are more likely to be accepted by society. But the existence of philosophy that has contributed to the development and understanding of civilization and learning itself is important. Mr. Hirai, what are your thoughts on practicality?
Mr. HiraiI think practicality is a matter of
time scale .
The talk of being useful within a few months or years makes the “time” to get answers from research the shorter the better, and the better it’s a necessary evil to avoid. However, if you set a goal (time limit) and aim for it, the mystery dealt with by the quest itself tends to be dwarfed according to the answer you can see now. The scale of the whole mystery becomes smaller.
It is often seen in the actual biological evolution that “it was not originally planned to be useful, but when the environment changed, it was expanded and became useful as a result.” It is not a practicality with a purpose set in advance, but a practicality that can be seen only as a hindsight at the end of an open time scale search. This means viewing time as something creative and positive rather than a necessary evil.
So, if research is aiming for something innovative, I think it will be necessary to have open discussions without speculating about the future.
Mr. TaniguchiIn the book I wrote, I first wrote a story about “Why use robots?” I think that the point is that the creation and use of AI and robots will increase the number of tools for ”
drawing out the mind .”
Technological change has accelerated the development of the field of science across all disciplines. For example, in brain science, a brain science boom has come with the development of measuring devices such as functional MRI. The same can be said for the history of the past. When technology is developed in every field of study, it is the timing when “human understanding and understanding of the world” will accelerate .
In that sense, there is no doubt that when philosophy develops, it is time for innovation to occur, and I believe that philosophy will deepen as technology advances.
Mutual understanding of each other’s fields is necessary
ーーWhat do you think is necessary to encourage exchanges between the two?
Mr. TaniguchiThere is an atmosphere in which AI researchers and philosophers are a little reserved with each other. Recently, apart from voluntary research such as academic conferences, there are many places for network-type research in which large-scale projects are discussed.
Such as ERATO and CREST . Among them, projects with clear goals such as technology development are often handled. Humanities researchers are sometimes involved in projects, but I sometimes feel that they are in a position of “bonus”.
In the first place, Japan does not allocate much of its budget to the humanities. I believe that the humanities department should also receive a large budget and focus on researcher education and long-term projects.
Also, the fact that it is not possible to philosophically evaluate research in science and engineering is another factor that prevents lively discussions.
Mr. HiraiIt’s not good to have a system that relies on someone like Taniguchi, who straddles science and engineering and philosophy, for discussions when creating a big movement in an organization. I think it would be good if more people were involved in various scales, and they were inspired by each other, leading to some new movements. I hope that AINOW’s efforts will serve as a catalyst for that.
Mr. TaniguchiIn order to be able to have such discussions with each other, science and technology people must have a certain degree of philosophical and ethical grounding. Merely memorizing Plato or “Thinking Reeds” in high school ethics is not enough.
I think it is necessary to have a broad and well-balanced foundation for learning about both science and engineering as well as philosophy.
Mr. HiraiI think the opposite is also true.
Even philosophers don’t have enough time to get to know the details of science. There is also the background of being too busy, but I think that the accuracy of the vision will differ depending on whether or not you have basic knowledge and understanding of methods.
In any case, understanding each other’s fields and having knowledge is necessary both for science and humanities. I think we are still at the stage where ideas to lower that hurdle should be unearthed.
Now is the time to learn philosophy
ーーWhat are the important points for people from other fields to learn about philosophy?
Mr. HiraiI think now is the best time to do some philosophy. If Aristotle and Descartes were living in modern times, I would be so excited that I wouldn’t be able to sleep at night (laughs). Because we stand on the horizon where various things are revealed every day.
Earlier, Mr. Taniguchi said, “Technology accelerates understanding.” Related to that, I think there are two kinds of understanding.
When a person faces a problem, he/she will feel satisfied and feel that he/she has “understood” by explaining and outputting “There are such merits” by putting the story into the purpose and means. However, this is not a true understanding, and there are many cases where “I just gave an answer within the given frame”. In other words, it means that the processing is done within the knowledge on hand.
On the other hand, it takes time to face the mystery itself, become familiar with the face, and advance the understanding to the point where you can understand the real part of it, so to speak, and it should be taken. It is not something that can be understood reductively by just putting the parts together. At that time, I think that technology will extend the limbs of our intelligence. We have to use it in practice and intervene in interactions. For that reason, the biological limbs (brain) that evolution has prepared for us are not enough, and we have to increase what we can do according to the mystery.
Concepts and techniques are extended bodies. Using that extended body, I challenge mysteries and practice every day. In the midst of that, I think that the mystery will finally be solved for us little by little. That is why I think it is important to spend time thinking about the environment in which research can be conducted and how to approach the research itself.
The development of AI is not only told by technological progress. Part of the foundation of the development of AI technology is the presence of a philosophy that has been approached from a wide range of perspectives such as intelligence, mind, and recognition.
“How to give emotions to AI?”
“What kind of impact and value will the created intelligence have on society?”
“What is intelligence in the first place?”
The questions that arise when creating such AI are themselves connected to philosophy. Also, the philosophy that explains what’s going on in people’s heads seems to be nothing special, but a very practical study.
By further deepening the dialogue between AI researchers and philosophers, it is expected that the foundation for international discussion will be created and that it will develop as a mutual academic field.