[Feature: Future Mobility Society] Roundtable: How Will Autonomous Driving Change Society?

Today, we are holding a roundtable discussion for our special feature, "The Future Mobility Society." Currently, the evolution and practical application of autonomous driving systems, as well as the transformation of automobiles due to energy shifts, are bringing about remarkable changes in cars and the society surrounding them.

In this context, we have invited people involved in technologies, services, and businesses related to autonomous driving today. In this roundtable, I would like to mainly discuss the current status of autonomous driving in Japan, its future possibilities, and its challenges.

Looking back at the history of autonomous driving, during the Cold War in the 1980s, autonomous vehicles using machine vision were developed due to military needs such as unmanned reconnaissance. In the 1990s after the Cold War, research on platooning for the main purpose of expanding traffic capacity, and the development and demonstration of autonomous driving that actively coordinates with roads, were promoted. From the late 1990s, the practical application of technology aimed at supporting drivers rather than autonomous driving began.

Entering the 2000s, research, development, and practical application of driver assistance technology were actively carried out, but there was no movement to actively promote the practical application of autonomous driving. Subsequently, movements toward the social introduction of autonomous driving overseas intensified, such as Google's declaration of autonomous vehicle development and the issuance of public road licenses for autonomous vehicles (USA, 2012). The catalyst for these overseas movements was also military needs. In fiscal 2001, the U.S. Congress mandated technology development to make one-third of ground combat vehicles unmanned by 2015, and based on that congressional mandate, DARPA (Defense Advanced Research Projects Agency) conducted races for autonomous vehicles.

Here, teams that developed systems capable of autonomous driving in complex environments through precise sensing of the surrounding environment and advanced information processing played an active role, creating the foundation for current autonomous driving technology. Google's autonomous vehicles follow this lineage. Pushed by these movements, study groups on autonomous driving were established in Japan as well, and autonomous driving systems began to be listed in the growth strategy announced by the government in 2013. Since then, various autonomous driving-related projects have been promoted.

Currently, the technological development and practical application of autonomous driving are proceeding in two streams. One is autonomous driving for private cars (owner cars). This is a stream of advancing driver assistance technology and increasing the things it can support. Currently, Level 3 autonomous driving functions, where the driver does not have to keep their eyes on the road while the system operates the steering, accelerator, and brakes under limited conditions, have been put into practical use.

The other is autonomous driving as a mobility service (autonomous driving for commercial vehicles). Based on the idea that autonomous driving is effective for securing means of moving people and transporting goods in a declining birthrate and aging society, and that practical application is relatively easy if the driving range and routes are limited, technological development and demonstration experiments for mobility services such as autonomous carts and buses, and transport by autonomous trucks, have come to be actively conducted. Overseas, mobility services using unmanned taxis have begun.

First, regarding the current status of autonomous driving, I would like to ask Mr. Kato, who is developing the technology, to speak, including the situation overseas.

The easiest way to understand the evolution of autonomous driving overseas is, after all, that Google-affiliated Waymo has started a completely unmanned autonomous taxi service in San Francisco. In addition, I think it's the release of Tesla's latest FSD V13, which requires a person to be on board but can drive almost automatically.

Waymo is for commercial/service cars and Tesla is for passenger/owner cars, but I feel they are both midway on the path toward ultimate autonomous driving. It's fair to say that the vision drawn globally about 10 years ago is becoming quite a reality. I think we've at least passed the halfway point, and since it has already been shown that fully autonomous driving is possible if you spend enough money, I believe we have reached the stage of lowering costs and increasing business viability.

Regarding mobility services, companies like Waymo are doing advanced work globally. It seems they have proven that autonomous driving is technically possible if you spend the money.

That leads to the question of whether it can be sustainably established as a business. Mr. Kato, do you think it will be viable as a business?

This depends on how you define the field of autonomous driving. Taxis and passenger cars are one of many areas, but the most difficult are indeed these taxis and passenger cars. This is because they drive in complex environments on general public roads.

Slightly before that, there are buses and shuttles. These run on public roads, but their routes are limited or their uses are restricted. Also, while trucks on highways have different difficulties, it is basically technology for driving straight, so it can handle many scenarios.

And when you go outside of public roads, there are various things like agricultural machinery, construction machinery, transport within factories, and large trailers. Problem-solving in these areas is progressing. Thinking that way, I believe business and economic viability can be achieved for mobility as a whole. However, I think we will find out from now on how much business and economy can be established when looking only at buses and taxis. Personally, I believe economic viability can be achieved in each area.

Mr. Aoyagi, what is your perspective as a taxi operator?

Overlapping with Mr. Kato's talk, the practical application of autonomous taxis is progressing in the U.S. and China. In complex situations close to those of general passenger cars, operators are emerging who are going beyond so-called demonstrations to provide commercial services that the general public can ride, albeit in limited areas and environments. I feel a possibility here first.

Just like general taxi services, everyone is paying fares to Waymo and other Chinese services. In the San Francisco ride-hailing market, Waymo's scale has surpassed the second-place Lyft and is approaching the first-place Uber. I think the fact that they are actually operating and succeeding as a business in regions where social acceptance has progressed is a major change in the last few years.

At newmo, which I manage, we are aiming for technical catch-up while collaborating with Mr. Omae and Tier IV on research and development in the autonomous driving field. At the same time, I manage a taxi company and currently operate about 1,000 taxis in Osaka, so we are at the stage of trying to see if the introduction of autonomous taxis can be done in Japanese public transport services by gathering various wisdom and experience from all over Japan. I personally feel great potential, and from a technical standpoint, I can see that it seems possible if a large investment is made.

After all, Japan will clearly face a labor shortage in the future. Not only in transport services like taxis and buses, but also in public services in general, including nursing care, the number of providers is clearly decreasing, so we must maintain the means of movement and the infrastructure of daily life. In that context, we need to automate as much as possible and suppress the competition for limited labor with other industries. How to coexist and live together with the public transport that is currently desperately supporting the infrastructure? How to form social acceptance before establishing it as a business model is the challenge we face as operators, but I believe there is great potential if we can overcome this.

Mr. Hashimoto, what do you think?

I joined AIST from Mr. Omae's laboratory and have been working there for about 20 years. Until last year, I was the team leader of the Mobility Service Research Team, but now I am the director of the Research Planning Office, looking at things a bit more broadly. Today, I would like to speak from a personal standpoint, not on behalf of AIST.

As for autonomous driving technology, as you said, I think it is being demonstrated that it can be done to some extent if you spend money. I think Tier IV is also leading the way in Japan by obtaining Level 4. Regarding autonomous driving mobility services in Japan, the government has set goals of about 50 locations in fiscal 2025 and over 100 locations in fiscal 2027, but the current situation is that implementation is not progressing as much as expected.

We have to think about why it's not progressing much, but when considering where autonomous driving services will first enter, I think there are two patterns. One is, as Mr. Aoyagi mentioned, things like "Robotaxis" entering places like the business districts of Osaka where there is demand, and I think it will truly become a competition. Since this is competition among private entities, the strong will win.

The other is in areas other than regional urban centers. I went to the U.S. the year before last, and in rural areas, there are regions where even Uber or Lyft won't come even if you call them. I think how to secure means of movement in such places will be a challenge. Furthermore, there is the driver shortage. Currently, bus services are being reduced in various regions within Japan. I think there is a possibility that autonomous driving will enter to solve the driver shortage for the sake of maintaining public transport. Regarding how to think about rural and urban areas, I believe the option of autonomous driving is an important keyword for both. I also hope to contribute to this field every day.

Many different perspectives have come up.

First, regarding the level of difficulty, as Mr. Kato mentioned earlier, there are levels such as trucks, buses that run on routes, taxis that run anywhere, and passenger cars where the driver goes anywhere at will.

Also, regarding rural and urban areas, it's obvious that rural areas won't be profitable, and while I think urban areas might manage somehow, on the other hand, there is also the fear of being swallowed up by foreign capital. Based on those points, Mr. Ioki, what do you think?

I am currently a faculty member at Keio's Graduate School of System Design and Management (SDM), a graduate school with a high ratio of working professionals. Here, rather than deepening a single specialized field, we are conscious of producing human resources who can solve comprehensive problems across multiple fields. We call this "horizontal expertise."

In my lectures, I teach Systems Engineering, which is a systematic way to design by perceiving the target as a system. Another thing I teach is how to think innovatively without being bound by precedent. As working professionals gain more experience, they tend to value precedent and their own experience, making it difficult to discard those and think of new things.

One of the things I say in my SDM lectures is not to mistake the "purpose" for the "means." It often happens that the purpose and means get swapped before you know it, and things become confusing. And once a purpose is decided, there are usually multiple means. In this case, if you proceed under the assumption that there is only one means, you will end up materializing it in a way that is unnecessarily difficult to solve. To avoid this, when designing something, I always have students think about what the higher-level purpose is and whether there are other options for the means to achieve it.

Actually, regarding the periphery of autonomous driving, I only came to know about it directly after joining a government committee chaired by Mr. Omae, so my years of experience in this field are not long compared to all of you.

I also have the conviction, of course, that pursuing autonomous driving is absolutely necessary for society. On the other hand, when autonomous driving is positioned as one of the means, I believe there are various options for the purpose of what kind of world to create and what kind of value to provide to whom. Thinking that way, I believe discussions arise about "to whom," "how," and "to what extent" autonomous driving should be provided.

Since I am not an insider, when looking from the outside, I am interested in the setting of the purpose: how will the world change by pursuing automation?

I feel there is an aspect where the extent to which autonomous driving technology is necessary in society is not yet properly communicated to the world. Slightly outside the logic that autonomous driving is indispensable because of the labor shortage, there must be a more important purpose that changes society, but perhaps that part is not being communicated to the general public. I am always thinking about that, but I haven't quite reached an answer. Today, I would like to hear the thoughts of the insiders on that matter.

What you said is very important. Indeed, autonomous driving now has many aspects and truly many means. In the past, autonomous driving was a bit simpler; for example, for military needs, there was talk of unmanned operation to ensure as few soldiers died as possible.

Or, since human driving can only allow about 2,000 vehicles per hour per lane, there was talk that if they drove in platoons with shortened inter-vehicle distances, traffic capacity could be increased by about three times. However, recently, purposes such as labor shortages or saving regional transport have also emerged, and various purposes are intertwined, perhaps making it unclear what autonomous driving is for.

Especially in rural areas, there is the idea of taking elderly people to the hospital with autonomous driving, but there is also the talk of making it possible to receive medical examinations without going to the hospital in the first place through telemedicine. Even for solving regional issues, there are approaches other than autonomous driving, so I think it is truly difficult to determine what is optimal.

Even when conducting demonstration experiments, there are times when it doesn't gain momentum for widespread use. It looks like a traveling circus troupe running buses here and there and then finishing. In profitable urban areas, I think even things like Waymo's Robotaxis have the potential to pay off, but are there any success stories of autonomous driving replacing public transport in places where the declining birthrate and aging population are progressing in rural areas?

Since it's the dawn of the era, I think it's still difficult on its own. I don't think there is yet a place where it has been established as a sustainably operable business and citizens are enjoying its benefits.

In innovation theory, the initial innovators are a few percent and early adopters are about 15 percent. Once you cross these early adopters, the economy finally starts to turn. Currently, autonomous driving has not reached 15 percent in any region or industry in the world. That's why subsidies are needed, but perhaps because few people can draw a picture of crossing the early adopters and entering the period of widespread use, few people think about providing subsidies.

As a result, it ends up tapering off. I think this is something that can be said for many things, not just urban mobility, but also transport within factories and highways.

The reason I decided to start newmo about two years ago was because I saw the current situation where regional public transport is becoming quite weak.

You can tell when you are a taxi operator, but even if people say "public transport is insufficient," in Tokyo and Osaka, operators are working hard while properly raising the salaries of drivers. In Tokyo, it's said that one can earn about 6 million yen a year, and in Osaka about 4.5 million yen, if they become a taxi driver, making it a job where one can properly make a living.

However, even looking at the same taxi industry, it has completely split into regions where taxi drivers are returning from the COVID-19 pandemic and regions where they haven't returned at all.

As for the order of the spread of autonomous driving, it might be easier to go from demonstration to social implementation starting from urban areas where there is population density and economic demand, but I think how to create that path will be questioned along with the raison d'être.

This year, Goldman Sachs wrote a report on robotaxis in China. Reading that, at this point, it feels like they are finally becoming viable as a business in what are called Tier 1 cities in China, such as Beijing, Shanghai, and Guangzhou. Currently, they are still operated with humans performing remote monitoring, with a limited number of vehicles, and in some situations, with people called safety drivers on board.

With this, it's hard to make the economics work, but according to future predictions, businesses will likely become viable in Tier 1 cities in the late 2020s. Furthermore, in predictions beyond that in the 2030s, if the price of autonomous vehicles, the necessity of safety drivers, and the efficiency of remote monitoring change, it will become viable even in cities called Tier 2 cities in China.

In China, the price of public transport services is about 3 dollars, which is cheaper than in Japan, but even so, it is thought it will be viable. I believe Japan can follow that path if we look at it in units of the next 10 years.

It is said that the problem of aging in Japan will become serious in the 2030s, and the providers of public transport will truly disappear nationwide. Like the example of China, I think the economics won't work at first, but if we overcome the hurdles in order, I believe we will be able to create a system to introduce autonomous vehicles to regions where even taxis and ride-sharing are already difficult. I feel that a path is starting to be seen little by little in what was previously darkness.

The point is that you should just do that with newmo. After all, you have a company that can do it. From that sense, I personally feel like there aren't many hurdles in Japan, but what are the challenges you are currently feeling?

Technologically, I think things are moving in a good direction. On the other hand, regarding the method of introduction, I think what you said earlier about "becoming a circus troupe" is exactly the challenge that is surfacing. The reason it becomes a circus troupe is that because there are no common specifications, standard models, or clear criteria yet, even if each region or individual tries to introduce it, there are people who can see it through and people who can't.

And the people who can't see it through move to the next place, and again move without seeing it through, which is why it becomes a circus troupe. If there were common specifications and standard models, instead of about 1,700 municipalities doing 1,700 different types of autonomous driving, it would settle into a few types of autonomous driving services. If that happens, I think the walls will disappear.

At this rate, efficiency is poor, so costs pile up. I believe that if common specifications and standard models are created, the field of mobility will progress significantly.

Regarding the institutional side, I understand that discussions are exactly progressing, so it's a timing where things are being sorted out.

Regarding the fact that it's not quite leaving the realm of demonstration, even Waymo has spent a considerable amount of time in San Francisco. I think there were various walls, but I believe there is a part where they accumulated the necessary amount of learning and overcame them.

Waymo and others, for example, have driven a cumulative total of about 100 million kilometers within the same urban area. It has been shown that when learning progresses in that way, they become able to handle more complex situations and various edge cases. I think there are many things that cannot be done by a single operator alone; there are parts where multiple operators compete, and I think it's necessary to learn together and gather that knowledge.

Conversely, regarding parts that can be done because it's Japan, would it be vehicles? Since the foundation of Japan's industry is indeed the automobile industry, I think we truly have to work as "All Japan" with each company taking a central role to steadily increase the number of autonomous vehicles, and I believe Japan has potential.

As for what is hindering it, I think it's everything; technology is one thing, but after all, I think the legal system is also not yet properly able to respond. Currently, relevant ministries and agencies are finally moving forward, but Japan has a tendency to act after seeing the situation in foreign countries, so I think it's a bit slow.

Even if you ask "To what extent should we respond?", it's difficult to get specific answers. The situation is that operators have no choice but to handle the fine details themselves. Japan is strict about safety and is not a society like the U.S. where you try it and if something happens, you make a rule or fix it, so I think it's a difficult place for autonomous driving.

China and the U.S. are similar, and it's often said that Japan and Europe are similar. In that sense, autonomous driving services haven't progressed much in Europe either, have they?

I think we have no choice but to do everything little by little—not just rules, but technology and, of course, acceptance. Actually, Chinese cars are already running in Japan, and Waymo is also in a situation where it's running in Tokyo. As Japan, I think it wouldn't be very good if everything eventually becomes foreign robotaxis, so I always hope that we can somehow prevent everything from being taken over.

Mr. Ioki, do you have anything regarding challenges?

There's no end to listing challenges; I think truly everything is a challenge. Even if the technology part is mostly solved and it's feasible by starting a small business and pushing forward, people keep saying after the fact, "Didn't you think of that?" when something happens.

It's probably not a world that can be settled just by coordinating properly in advance and getting people to agree before starting, so I think we have to keep doing an iterative cycle of trying it and dealing with it if something happens, but eventually, we will hit a massive wall. To put it bluntly, the day will absolutely come when the first serious personal injury accident involving an autonomous vehicle occurs in Japan; it's unavoidable.

Regarding whether we can prepare anything for that, I think insurance companies are thinking about various things in terms of covering it with insurance. On the other hand, regarding social acceptance, I think various activities are being done, but I don't think social acceptance will progress simply by disseminating information for something that doesn't exist yet, so it's necessary to have people see what happens by actually running it.

If so, there's no choice but for the operators to make up their minds and move forward, but at that time, who will provide support and who will act as the advocate for moving forward? I wish there were someone with the role of protecting when something happens. Otherwise, I feel like we will hit a wall somewhere and it will be a disaster.

For example, when the U.S. was pursuing manned spaceflight, it was factored in from the beginning that people dying was unavoidable. At the Kennedy Space Center in the U.S., there is a huge monument where the names of those who died are written. There are 25 names written there, but a large space still remains. In this way, while making the maximum effort to ensure safety, the intention to do it with a certain degree of resolve is expressed in the world of U.S. manned space development.

I'm not necessarily saying autonomous driving is the same, but while steadily advancing development toward a bright future, it's hard to touch on bad stories. I think there could be more talk about "who will protect how when something happens." This topic is very difficult to bring up, and if operators can't even bring it up, I feel like someone else has to say it.

Are you saying that society should have a bit more tolerance? I think that's true, but I also think it might be that people don't think the world is that inconvenient yet. While saying oil will run out, everyone is wasting it. I wonder if there will be a response when it truly runs out.

Society says autonomous driving is "necessary," but because the situation hasn't become that serious yet, there is an atmosphere where if there is an accident with autonomous driving, even a minor property damage accident is strictly scrutinized. I thought that if the situation becomes such that it's truly necessary and the world can't function without it, things might be a bit different.

But we shouldn't wait until then, should we?

That's true.

Now, the next topic. Currently, autonomous driving is loaded with a large amount of AI and GPUs. Somehow, an atmosphere is being created now that high-performance AI is absolutely necessary for autonomous driving, but I feel like we are being pulled into the playing field of companies that are good at those things.

In my personal view, it seems Level 4 could be achieved without using AI, but what do you think, Mr. Kato?

Autonomous driving is a broad field. If you look only at the difficult parts, investment is required, so the return is small. However, looking at it as a whole, there are quite a lot of things that can be done even with current technology.

For example, the price of a single tank truck in a steelworks is very high, so it's viable even if you add an autonomous driving system costing several million yen. Also, since the main task inside a factory is basically patrolling, the necessity for high-performance AI is limited. In contrast, taxis and general passenger cars have difficulties. If you put a several-million-yen autonomous driving system on a several-million-yen vehicle, there is a price impact right there, and furthermore, more difficult driving is required.

I think high-performance AI is indeed necessary in those fields. I feel many people are starting to have a common understanding recently, but the challenge of autonomous driving is to challenge the trade-off between safety and availability. Since Tier IV has obtained Level 4 certification/permission, we have become able to reliably avoid collisions under specific conditions.

However, when this is deployed in the real world, there are scenes where it stops moving. Because it is designed not to collide, if there is even a slight obstacle, it takes the option of "not moving." The introduction of AI that is being talked about now is for judging that it's "okay to move a little" in these non-moving situations. If AI judges based on human common sense that "it's crossing the white line, but it's okay to go," it becomes able to move. There is no doubt that having AI makes it easier to solve the trade-off between safety and availability.

Even with current technology, we can guarantee that it won't collide, but when it comes to whether it can drive smoothly, there are many difficult parts. This doesn't necessarily have to be AI; sometimes humans help remotely, or the infrastructure side helps. In this way, it is already possible to perform safe driving, so I think AI is one means to further increase availability.

You've given us wonderful insight. In other words, it's not technology to stop the car, but technology to make the car proceed. This is something only someone who has developed autonomous driving would understand. If you do it normally, it keeps stopping and becomes useless, so AI is what gives the go-ahead.

I apologize for only using Waymo as an example, but what Waymo has done is capital investment into creating service cars that combine not just camera vision but various sensors. It's about utilizing computers with high-performance GPUs to solve cases that couldn't be solved before by combining them.

However, they have invested about 1 trillion yen in total. Major Chinese players who have followed in a way that mimics this have also invested on a scale of several hundred billion yen so far.

In contrast, there are two things I feel potential for in the utilization of AI. In actual urban traffic, when congestion occurs, there is a judgment of which route to take, and there are times when actions to take when an accident occurs, such as for emergency vehicles, are not anticipated. Also, in Japan, there are roads where a lot of construction is being done. I think the utilization of AI can supplement conventional methods in learning these things.

Another point is whether 1 trillion yen is really necessary. By increasing the speed of development and actual commercialization, parts that can be repurposed may have emerged. This is exactly the benefit of AI utilization progressing in various fields in the 2020s; while the early players in the U.S. and China did it with a kind of capital power, because it is a latecomer, Japan might be able to make investments that utilize the benefits of AI. Thinking that we might be able to take the path to practical application quickly and efficiently, I feel that the acceleration of learning by AI has emerged as a dramatic possibility.

I think it's a matter of commoditization. This is a huge opportunity. As Mr. Aoyagi mentioned, the companies that started first invested 1 trillion yen. Companies that started just before commoditization invested 100 billion yen. If you start after commoditization, it could be that "it didn't even cost 10 billion yen." It would be great if Japan could win here.

Now, when people hear "AI," many think of generative AI, but there are three types of AI. There is traditional rule-based AI, which operates based on rules and conditions set in advance by humans, predictive AI, and generative AI. Predictive AI didn't exist when TIER IV started autonomous driving. The term "deep learning" didn't even exist. Because of that, we could barely even detect objects. However, now that predictive AI has become commoditized, we can do this without spending much money.

And now, since various companies have spent trillions of yen on generative AI, many models have emerged. I think the winning strategy from here is to move forward while keeping costs down.

So you're saying there are actually more advantageous aspects.

Another thing I wanted to ask is that back in the days of 8-bit PCs, there were many different types of computers, but before we knew it, they all became Windows machines. There were various mobile terminals as well, but now it has become a choice between iPhone and Android. Geographic information services have also mostly been swallowed up by Google Maps. At first, many different people are making various efforts, but in the end, there is a sense that they are swallowed up by giant overseas companies.

However, rather than saying that's outrageous, I feel that Japanese people are able to use high-quality services at a low cost, so it's not like they are becoming unhappy. Even in autonomous driving, if Waymo and others come into Japan and dominate, people might just think, "It's great that we can ride taxis cheaply now," and a day might come when we use it as if it were perfectly natural.

However, since that would be unfortunate for the Japanese people, there is also the idea that Japan must work hard as a nation. What are your thoughts on that?

I think autonomous driving can be viewed in many ways. What does it look like in terms of social value, industrial value, and academic value? I think the perspective probably differs for each.

When looking at social value, if users can fully enjoy safety and availability, there is a possibility that it doesn't necessarily have to be a domestic autonomous driving service. However, as a risk, there is a possibility of falling into a digital deficit ten years from now.

Probably the biggest one is industrial value. The reason many companies are trying to do autonomous driving is because it becomes an industry. Therefore, if we cannot produce industrial value, there is a possibility that the Japanese automobile industry will just become a manufacturer of empty shells.

Academic value is something anyone can work on. Since autonomous driving is a representative example of physical AI, I think researchers are enjoying their research.

As with the examples of PCs and smartphones, I think something similar will happen in the field of autonomous driving. I believe there is a possibility that Japan can produce open-source software like Linux, Android, or DeepSeek.

In fact, your "Autoware" is exactly that, isn't it?

The "Autoware" we are developing is trying to take exactly that position.

I believe that incorporating various overseas technologies is essential for moving this field forward.

At the same time, autonomous driving is established through a combination of various players and various technologies. Waymo, which we've been talking about, uses Jaguar vehicles, and there is room for Japanese industry to be involved there. Regarding sensors, Japanese companies like Sony are handling the most difficult parts. Just as iPhones have Japanese-made parts, there is the question of how we fulfill our role within that.

Furthermore, the themes of recent unstable international situations, economic security, data preservation, and security are inseparable from this field of autonomous driving transportation. I think it is important to have sustainable options domestically as well.

Having alternative options or not depending on a specific country leads to peace of mind. I think increasing options ultimately overlaps with the interests of general users.

Since Japan has a strong automobile industry, I think there are parts of this that can be turned into an export industry. If we don't do it, everything will be swallowed up, so we definitely must do it.

I think there are two different perspectives on this. One is the digital deficit, but another is that once a market share is taken, it's difficult to take it back. This is true for overturning Windows, and for iPhone and Google Maps as well. So once it reaches the level of Windows, if you ask if we are going to make something to replace Windows now, it's technically quite difficult.

Also, if one thing monopolizes, you have no choice but to do business there, and I think it becomes even tougher in terms of cost.

On the other hand, based on what Mr. Ioki said at the beginning, if we consider autonomous driving as a means in the first place, and speak of it as a means to make everyone's mobility happy without considering the problem of digital deficit, it's a dilemma because it seems like cheaper overseas options might be better.

In Japan, we have the concept of "Japanese spirit with Western learning" (wakon-yosai). Exactly when Keio University was founded, the country opened up and overseas systems and technologies came in, but I think new Japanese entrepreneurs designed a significant portion of it to create the foundation of modern Japan. So, whether from the standpoint of a business person or in terms of how the nation should be, I think it's necessary to consider how we contribute.

While maintaining that intention, it's not about closing ourselves off tightly, but rather incorporating overseas technology while also making it our own technology.

Mr. Ioki, do you have anything to add?

While the highest priority is placed on how to realize autonomous driving, if we talk about the layer outside of that, I think people's lives and needs will change as technology gradually improves in perfection and spreads. This means that society changes while both move together.

There was talk about how autonomous driving is needed because elderly people in rural areas are struggling because they can't go to the hospital. If it takes, for example, nearly five years for that technology to actually be used, how will society and people's lives change during that time? Also, how will the various technologies surrounding people's lives change? If society is created while both move in parallel, I think we should intentionally design a world that is more cooperative toward the realization of autonomous driving. Shouldn't there be a forum for discussion with the parties involved about how society can change to be more cooperative toward autonomous driving?

Regarding the logistics issue, there is no doubt that there is a shortage of truck and taxi drivers, but it can also be said that current logistics depends too much on mobility. How much of a logistics burden do we impose just to buy one bottle of tea? In that sense, how the logistics system will change in the future under the influence of affinity with automation is something that should be discussed on the autonomous driving side as well, and there could be talk like, "If the logistics side is going to do this much, then we don't have to do that much here." I think the situation might change a bit if a worldview where both sides meet halfway emerges more in the discussion of autonomous driving.

Currently, discussions about autonomous driving are repeatedly focused on whether it's okay even in extremely extreme cases. Because of that, it has become necessary to equip many sensors and perform a lot of processing, but I wonder if there might be some missing discussion there.

Listening to today's talk, I felt that there are parts where discussion needs to progress more in the mutual relationship between society and autonomous driving.

Needs on the side of society are like a process of alignment, but do those on the technology and business side have anything they are doing to raise awareness toward widespread adoption in society?

Although it's limited to the field of Osaka, when talking with local governments and businesses in the Kansai region, the conversation goes beyond just autonomous driving software and vehicles to how the transportation infrastructure and operators can contribute to society. Themes such as how to allow passengers to get on and off safely, unlike a bus, or how to coexist in traffic jams, have been tested quite a bit, so we are now able to have specific discussions.

As a business that entered recently like us, I feel that a base for accumulating such discussions has been created, and I think that is exactly what Mr. Kato and predecessors in Japan have opened up.

Each local government in Osaka is well aware of the post-Expo period and is thinking carefully about how to utilize technology, so I actually feel that Japan might have more local governments and players that can collaborate well.

Does it often turn into that kind of discussion?

I'm talking with Sakai City and Osaka City, and everyone is quite fast; we are able to have surprisingly specific talks about specific routes and areas. It varies by region, but that's how it is in Osaka.

Since ideals and society are highly abstract topics, I think it's important how much we can abstract them as concepts. In terms of the value of autonomous driving, there are two ways of thinking: forecasting (predicting the future) and backcasting (thinking about the present by calculating backward from the future). What I think of in terms of forecasting for autonomous driving is that "it must be able to do what humans cannot do." For example, humans cannot absolutely avoid collisions. Humans cannot drive at all times.

Doing what humans cannot do is highly valuable, so from the standpoint of a university faculty member thinking in terms of forecasting, it means realizing what humans cannot do, and then having companies and others think about how to use it.

Backcasting that as a business operator, the ideal would be for it to be available "anytime, anywhere, anyone." It's a service that can run 24 hours a day, and can be expected to function in any environment. Also, "anyone" means, for example, if my own grandmother were a bus driver, I think that would be an amazing society.

Therefore, if we can realize what humans cannot do, I think the destination will be a service that fulfills "anytime, anywhere, anyone."

I think it's important for young people to think more and more about the future. Saving the mobility of the elderly in rural areas is very important, but for young people, it feels like someone else's problem, and I don't think there are many opportunities to think about it together.

I always think it would be great to consider a society where there is mobility that excites young people.

To be honest, I myself sometimes think that autonomous driving isn't really that necessary (laughs), but what I learned this time is the importance of autonomous driving being a means and what you set as the goal. It's not just about solving the driver shortage; there are various things to consider, such as Japan's national strength. I think there are parts where everyone thinks about it too much in a jumble and gets hesitant.

Personally, I think it's better to use machines not just for unmanned operation, but to help people by collaborating with them, so that people can continue to work and be active in society even as they get older. However, after talking with everyone today, I feel that the direction of autonomous driving is bright, that it is indeed necessary, and that it's better to put it into practical use.

Thank you very much for your time today.