Skip to main content

This page has been translated using TexTra by NICT. Please note that the translation may not be completely accurate.
If you find any mistranslations, we appreciate your feedback on the "Request form for improving the automatic translation ".

Third meeting of the Sub-Working Group on Social Rules for automated driving vehicles in age of AI

Overview

  • Date and Time: Tuesday, February 27, 2024, from 10:00 to 12:00
  • Location: Online
  • Agenda:
    1. Opening
    2. Content
      1. Explanation by the Secretariat (Main opinions of the Second age of AI Sub-Working Group on the Review of Social Rules for automated driving vehicles and Possible Issues Based on Your Opinions)
      2. How to Understand the Security and "Defects" of the automated driving Program
      3. JAMA's Reflection on the Discussion of Liability for Traffic Accidents Involving automated driving vehicles
      4. Current status, etc. of the crime of negligence in the pursuit of social activities causing death or injury
      5. The introduction of the criminal immunity system does not gain the understanding of the people (especially victims)
    3. Exchange of opinions
    4. Adjournment

Materials

Minutes

Councilor Kodama: First of all, I would like to make an administrative contact. Today's meeting will be held completely online. Members are requested to turn on the camera during the meeting and unmute the microphone when they speak. In addition, I would like to ask you to mute the microphone when other people are speaking. In addition, spectators are requested to turn off the camera and microphone.

Next, I would like to check the materials. As stated in the agenda sent in advance, the materials include the agenda, the list of members, the secretariat explanatory material, the material explaining Mr. Fujita, the material explaining Mr. Hatano, the material explaining the Ministry of Justice, the material explaining Mr. Takahashi, the material submitted by the members, and the list of attendees. If you do not have them, please contact the Teams chat function or the secretariat by email. As for the attendees today, due to time constraints, I would like to distribute the list of attendees in your possession instead of introducing them. In addition, I heard that Mr. Harada is scheduled to leave the meeting in the middle.

In addition, Mr. Kamiya, Legal Advisor, Mr. Ozawa, Representative Director, Mr. Matsunaga, Deputy Representative Director of the Association of Crime Bereaved Families of the Seki Higashi Kotsu (Ai no Kai), and Mr. Ohno and Mr. Suganuma from DADC, who was announced last time, are participating as guests today. The people I just mentioned are scheduled to continue to attend next time. Please be aware that the materials and minutes of the plenary meeting will be made public at a later date.
Then, I would like to ask Mr. Kozuka to proceed from here. Mr. Kozuka, please.

Mr. Kozuka: Thank you very much, Mr. :

I will proceed according to the agenda. First, let's move on to agenda 2 - (1). As an explanation by the secretariat, I would like to ask you to explain the main opinions of the previous Sub-Working Group and the expected issues based on those opinions. Then, Counselor Suga, thank you very much.

Counselor Suga:
(Hereinafter, explanation will be made based on "Material 3: Third Secretariat Explanatory Material")
Material 3: Based on the explanatory material of the 3rd Secretariat, we explained the main opinions at the 2nd SWG and expected issues based on the opinions at the 2nd SWG.

As I do not have any materials from here, I would like to speak orally. In 2022, there were 2,550 fatal traffic accidents, of which 1,971, or 77%, were caused by human factors due to driver errors. Among these 2,550 accidents, accidents such as those between motorcycles and pedestrians are included, so cars are not necessarily included in the denominator, and the actual number may be more than 77%. In any case, it is clear from this data that many fatal traffic accidents are caused by driver errors. automated driving systems, particularly unmanned automated driving systems, substitute for the recognition, prediction, judgment, and operation performed by drivers. Therefore, if safe automated driving systems are widely spread in the future, it is expected that the majority of traffic accidents caused by driver errors will be reduced. To that end, I would like to ask for discussions in the direction of enhancing safety. That is all from the Secretariat.

Mr. Kozuka: Thank you very much, Mr. . You organized various points of discussion. Finally, you pointed out that the problem is how to spread technology that is beneficial to society. In relation to that, I would like to ask about Agenda 2 - (2) How to understand the security and defects of the automated driving Program. Mr. Fujita of the University of Tokyo, please.

Fujita Member:
(Explanation based on "Material 4: How to Understand the Security and' Defects' of the automated driving Program")

Best regards Now, I would like to make a report on how to understand the security and "defects" of the automated driving Program. At the first meeting, I said that it is necessary to organize the basic concept of safety required when operation is managed by a system without the involvement of a human driver. Today, I would like to report on my response. The main purpose of this report is not to assert my opinion strongly, but to use it as a reference for your future discussions. The secretariat has just mentioned the clarification of the defect concept, but this report does not clarify the defect concept. I think the manufacturer would like to hear about the specific extent to which safety should be ensured, but such a matter is not clear. However, as a major premise, I would like to talk to you today because I am concerned that the basic awareness of how to think about defects is not shared. At the same time, I would like to briefly point out at the end that the content of the program will be in Issue. Under the condition that human drivers are allowed to operate automobiles depending on the program, which should be partially the case not only at Levels 4 and 5 but also at Level 3, there is a possibility that the evaluation of the security of the program will be meaningful as an interpretation of the current law and legislative theory when considering the ex post facto liability of accidents caused by automated driving vehicles. Please take the next page of the slide.

As a starting point, no matter how good a automated driving program is, it will not eliminate all accidents. Conversely, the very existence of an unpreventable incident does not immediately mean that the automated driving Program is unsafe. I don't think anyone would disagree on this point. I also think there is probably no objection to the point that the automated driving program is required to be able to provide at least safer driving than the average human driver, and if this is not possible, it is not safe. And if all the accidents that the program can't prevent can't be prevented even by human drivers, then the program can drive safer than humans. The problem is that most accidents that can be prevented by human drivers can be prevented, and many accidents that cannot be prevented by human drivers can also be prevented, but it is possible, albeit with a very low probability, that accidents that can be prevented by human drivers cannot be prevented. The question is how to evaluate such programs. The following discussion will be based on the premise that such a situation is possible. If the program design side can declare that all accidents that can be prevented by human drivers can be prevented by the program, and if such accidents occur, we can proceed with the discussion on the premise that we will not complain even if it is said to be defective, I think there is no problem in not paying attention to the future discussion. However, as a premise that such a situation is possible, I would like to explain two different views on the premise of a very simplified setting. Please take the next page of the slide.

I will not read it out loud due to time constraints, but in order not to complicate the discussion unnecessarily, I will presuppose the content written on the slide on page 3. In the first place, I would like to speak on the premise that there is a legal system in which the presence or absence of liability depends on the presence or absence of defects and safety, that there are no situations in which user negligence should be a problem, that there are no simple program bugs, etc., and that there are no restrictions on programs other than safety. Please show me the next page of the slide.

If we talk about it too abstractly, it will be a dogfight, so I will talk about the liability of the operator under the Automobile Liability Act as an example. Assume that a level 4 automated driving vehicles has caused an accident within the scope of ODD. Under the current law, the operator is not exempt from liability unless he / she proves the three conditions for exemption under Article 3 of the Automobile Liability Act. However, in a level 4 automated driving, the operator does not need to do anything in ODD, so basically, the driver's negligence in driving is not a problem. In that case, whether or not the operator is exempt from liability is determined by the presence or absence of structural defects, etc. under the current law. Conventionally, whether or not the operator was negligent in his / her act under the specific context of the operating environment immediately before the accident was questioned, but in automated driving, it is not a problem, and the presence or absence of structural defects, etc. is the determining factor of the liability. Please take the next page of the slide.

Based on this, an ex post facto and individual evaluation approach can be considered as one way of thinking about determining the presence or absence of structural defects. In this way of thinking, in parallel with the determination of the fault of the human driver, the control performed by the automated driving program is evaluated from the viewpoint of whether or not the control would have been determined to be negligent if the human driver had performed such an operation. In relation to the said accident, if the program has performed the control that would have been determined to be negligent if the human driver had performed such an operation, there are structural defects, etc. due to the lack of safety, and therefore, it is an approach that the vehicle operator is considered to be responsible. Another is the idea of the ex ante and general evaluation approach, in which the extent to which accidents can be prevented by the program is determined in advance. As an ex ante probabilistic judgment, if the operation control of the vehicle in a specific accident phase can provide significantly safer driving than the human driver (although the question of how safe it needs to be remains), even if the operation control of the vehicle in a specific accident phase is determined to be negligent if the human driver had performed such an operation, it is not immediately considered to be defective. Please go to the next page of the slide.

In the ex post facto and individual evaluation approach, instead of the driver's negligence becoming a problem, the way of judging the driver's negligence is brought into the structural defect part as it is. In that sense, the place of the judgment changes, but the same content is judged, so it is a status quo, so at first glance, it seems to be a safe interpretation. However, in this way of thinking, there is a problem that the driver says that he or she may depend on the program, but on the other hand, it is a problem that the driver is required to guarantee a certain content to the operator regarding the driving control realized by the program. It is still understandable if a local bus operator is assumed as the operator, but if an individual driver is assumed as the operator, it is more questionable. Please go to the next page of the slide.

On the other hand, if the ex post facto and individual evaluation approach is abandoned and the ex ante and general evaluation approach is taken, the problem on the part of the operator will disappear. However, from the perspective of the victims, if they are damaged by driving that a human driver can hold accountable for their fault, there is a possibility that there is no one to hold them accountable, and there is a possibility that no one will be held accountable. There is a possibility that the driving will be considered to be negligent by a human driver, but at least the person who is driving the car will not be held accountable, so there is a possibility that the feelings of the victims will not be satisfied. The biggest problem in taking this approach is that this problem, and by extension, automated driving may lack social acceptance. Please take the next page of the slide.

It is not the purpose of today's report to discuss theoretically and thoroughly which idea is correct, but I would like to make a brief comment. First, I think that it must be said that the determination of structural defects, etc. by the ex post facto and individual assessment approach is at least problematic as a form of liability for vehicle users under the Automobile Liability Act. It may be an obstacle to the use of automated driving, and I think it will be particularly noticeable when users other than business operators are assumed. Second, in the case of taking the ex-ante and general assessment approach, some legislative measures may be necessary, but as described in the third point of the material, in relation to the protection of victims, it is sufficiently considered that some kind of supplement is provided at the same time. For example, even in the case where there is no structural defect and no liability for vehicle users occurs, it is not impossible to consider taking measures such as providing a minimum guarantee at the same level as mandatory automobile liability insurance in the form of an automobile damage compensation project under Article 72 of the Automobile Liability Act, which is currently performed in the case of an uninsured hit-and-run accident. Specific requirements, etc. need to be considered separately, but if social acceptability can be enhanced by adding such measures, it can be one of the options. Please take the next page of the slide.

As mentioned above, the existence of defects can be mentioned as the cause of liability under the Product Liability Act. There are common problems between structural defects under the Product Liability Act and defects under the Product Liability Act. Please note that it is an established idea that the presence or absence of defects under the Product Liability Act is determined at the time of delivery, but this does not mean that the determination of defects under the Product Liability Act is necessarily made by the ex ante and general evaluation approach. This is because it is at least logically and conceptually possible to determine the safety of the program at the time of delivery by the ex post facto and individual evaluation approach. Of course, it is logically and conceptually possible and not contradictory, and whether it is appropriate or not is a completely different matter. Here, I would like to proceed by stating that the same problem exists under the Product Liability Act as the problem of structural defects under the Product Liability Act. Please show me the next page of the slide.

Criminal liability is not my specialty, so I will not go into it too deeply. For example, in the Automobile Driving Punishment Act, a defect itself is not a constituent feature of an automobile. Therefore, unlike in the Automobile Liability Act and the Products Liability Act, the way in which a defect is perceived does not immediately affect whether or not a person is liable. However, if it is interpreted that entrusting driving to a automated driving program that lacks safety means that a person has "failed to take necessary care in driving an automobile" as referred to in Article 5 of the Automobile Driving Punishment Act, whether or not a defect is present, and thus the way in which a defect is perceived, may have an effect. In addition, in relation to the liability of a manufacturer, whether or not a program lacks safety may be a problem in a slightly more direct form. However, in the case of criminal liability, recognition or recognizability of a defect is required, and in that sense, it will be different from the case of civil liability in that safety or the way in which a defect is perceived is not a direct consideration that immediately determines whether or not a person is liable. Please go to the next page of the slide.

So far, I have explained two opposing views on how to understand security in the case where liability legislation, in which the presence or absence of liability is determined depending on the defect, is maintained. However, I think it is necessary to consider the following problems, which have been neglected this time, at some stage. I will omit details, but there are some difficult questions about the concept of defects in learning and evolving programs, in particular, how to organize the concept with the concept that the time of delivery is used as the standard of judgment under the Product Liability Act, and whether liability may occur if users fail to fulfill their obligations such as updating, but whether to consider it as a defect or a separate component of negligence, and how to consider the causal relationship between defects and damage, especially when taking a preliminary and general evaluation approach. In addition to that, it is also necessary to consider whether ethical restrictions on the content of the program, typically referred to as the trolley problem, are ethical or legal, and if so, what is the meaning of it, for example, whether liability for accidents has any meaning. All of these seem to be long-term Issue, but as a starting point, I have mentioned them so that you can recognize the tasks to be considered. That is all.

Mr. Kozuka: Thank you very much, Mr. . You have raised a very big issue, and I think we can discuss it in depth, but I will move forward with the process.

Member Hatano will make a presentation as JAMA's consideration of the discussion on the responsibility of the traffic accidents involving automated driving vehicles in Agenda 2 - (3). Thank you.

Hatano Member:
(The following explanation is based on "Material 5: JAMA's Consideration of the Discussion on the Responsibility for Traffic Accidents Involving automated driving vehicles")

Thank you for your introduction. I am Hatano of Honda-shi, Chief Inspector of the automated driving Task Force of the Japan Automobile Manufacturers Association. Now, as the title says, I would like to make a presentation on the considerations of the Japan Automobile Manufacturers Association. Please turn to the next page.

This is a reprint of the material presented at the first Sub-Working Group as a study material for the social rules that JAMA is considering. In particular, the illustration on the left shows the concept of the concept of what functions, roles, and responsibilities will be used in the area that automated driving intends to provide services, and how to respond in a trinity including infrastructure, cooperative systems, and traffic participants. I thought this figure was a little difficult to understand, so on the next page, I have supplemented the part on how this can be derived.

In automated driving, specifications are not usually determined from the beginning, and the project starts after several stages: functional demonstration, functional demonstration on the road, demonstration of unmanned driving and service demonstration. Initially, the scope of services to be provided is shown in the figure on the left, and consideration begins with the inclusion of risks that are difficult to foresee and risks that are difficult to avoid. In general, the automated driving system is not perfect, so it is normal that the area where we can observe the traffic rules and ensure safety is generally smaller than the expected service area. There are several approaches to ensure safety, but as a representative example, it is a general idea to narrow down the area to be provided so that the automated driving system can respond. Even so, if the performance of the system is not sufficient in the service area, as shown in the figure on the right, we will take measures to improve the infrastructure in some cases, such as installing crosswalks, traffic lights, right turn lanes, guardrails, and responding to cooperative systems. In addition, although it depends on the relationship with the surrounding residents, it is assumed that the automated driving will be able to operate safely in the service area that is provided with peace of mind for the first time after arranging the relationship with people, such as ensuring that the traffic participants in the surrounding area who share the traffic environment observe the rules. Therefore, operating a automated driving does not necessarily mean that unexpected events are present everywhere, and it is extremely important to prepare to ensure safety by cooperating with relevant parties in advance. Next page, please.

Based on this preparation, we have organized the perspective of how to ensure safety. The figure on the left of the document is repeated. It is a figure classified into four areas based on the Safety Technology Guideline of automated driving, which was presented by the Ministry of Land, Infrastructure, Transport and Tourism, and the guideline that reasonably foreseeable and preventable accidents resulting in injury or death will not occur. As I mentioned earlier, the area where safety measures have been taken in the Trinity is, in the figure on the left, the blue area where it is stated that design and manufacturing must be done so that accidents do not occur. I think you can understand that even if there are no defects or errors, the number of avoidable, incidental, and unavoidable accidents is not zero if the automated driving is operated beyond the range in which safety is guaranteed in advance. In order to ensure maximum safety in the areas we intend to provide, we believe that it is extremely important for the manufacturer to clarify how to identify, examine, and evaluate the blue areas. As described in No. 1 in the yellow area at the bottom of the document, we believe that it is important to make reasonably foreseeable accident cases finite, that is, to arrange them so that a validation can be made, to share the criteria for performance that can be avoided, including the testing method, and to use the results of this as a baseline for forming an agreement with society. However, as I mentioned earlier, it is also true that the number of cases in which it is difficult to avoid and an accident occurs even if there is no defect or negligence is not necessarily zero within the predictable range. It is also necessary to establish standards for how society will judge and accept such cases. The third part of the material is about cases that cannot be foreseen at all in advance. In this case, we could only consider responses ex post facto, so we thought it was important to discuss ex post facto treatment in advance, including that point, so we introduced it as an expected value at the first subworking. Next page, please.

It shows examples of factors in which accidents are assumed in specific cases. If accidents occur in the area where design and manufacturing are performed so as not to cause accidents, which is the blue part on the document, Factor 2-1 is considered to be poor recognition and judgment that automated driving vehicles cannot correctly recognize the driving condition and cannot avoid accidents. There may also be situations where sensors or other systems are defective or out of order and cannot be detected in advance. It is also assumed that there is a problem in recognizing the location of the vehicle on a map, etc., or that the information from the infrastructure, etc. is inappropriate. In addition, considering Factors 2-2, it is assumed that the automated driving vehicles is not properly controlling the vehicles, that is, poor judgment and action. If there is a defect or failure in the functions such as running, turning, and stopping, or if the travel trajectory is not calculated properly, computer problems are also assumed. We would like you to understand that these events should be design and manufactured so as not to lead to accidents, and even if there are defects or malfunctions, they are arranged on the premise that they will be discovered and dealt with by the automobile manufacturer or the party who uses and operates them before they lead to accidents. Next, in Factor 1, there is a blue area that ensures safety as much as possible, but unfortunately, automated driving cannot be avoided. Even if there is no failure or defect, for example, there may be a situation in which an obstacle that is too close to the automated driving vehicles to be avoided suddenly blocks the track. In other words, I think the situation will be beyond the physical limit, but for example, if an oncoming car suddenly runs in front of the vehicle, or if pedestrians or other nearby traffic participants suddenly rush out at a distance that cannot be avoided, I think it is physically difficult to respond. In addition, as Factor 1-2, although we are prepared to respond as much as possible, malicious attacks such as cyber attacks are an area where it is quite difficult to respond in advance. If these things interfere with proper driving, falsification of the speed limit, wrong lighting of the signal, or wrong recognition may lead to an accident. However, we recognize that it is difficult to reasonably guarantee all cases in this yellow area with current science and technology. In this regard, it is indicated that it is necessary not only to assess the technical limitations, but also to organize the concept of how all traffic participants should follow the rules, and to hold comprehensive discussions, including the possibility of pursuing criminal liability. For your reference, it is difficult to respond in advance to cases that cannot be predicted. However, in the case of an event that cannot be predicted at all, such as the destruction of a road by a natural disaster or the sudden fall of something from somewhere, it is necessary to discuss how to handle the event in advance because the response will be ex post facto. Next page, please.

In this way, the manufacturing and industry sides are doing their utmost to ensure safety, but it is assumed that various parties are involved in the actual accident. Manufacturers and business operators try to provide services, but they receive a certain examination and start using and operating them with permission. Naturally, business operators and owners maintain them in a planned manner, and vehicles are operated in a good traffic environment. When a person drives a car, it is customary for the driver to fulfill his or her responsibilities within the scope of the operator. However, as shown in (1) on the document, if there are areas where the responsibility differs between accidents caused by a driver and accidents caused by automated driving vehicles, I think it is necessary to clarify them. This is in line with what Mr. Fujita mentioned earlier. As long as vehicles are properly design, manufactured and used, accidents are basically expected to be avoided. If the subject of liability is to be significantly changed from the past, we believe that it should be advanced after confirming the legal basis and the premise of the discussion. I would like to touch on the actual cases that have come up until the last time because we had a discussion today. Of course, we are aware of past cases concerning defects, but basically, automobile manufacturers design and manufacture vehicles so that they satisfy the normal safety standards, have them appropriately inspected, and sell or use their products. Even with such efforts, accidents that are unavoidable will be unexpected, but I think it is necessary to reconfirm the extent to which we should bear responsibility for them. In addition, in area ③, we are aware of cases of serious accidents in the transportation business. We understand that the security of the project is usually secured and permitted not only by the security of the vehicles but also by the transportation business operator's education to the relevant parties, as well as by the inspection of the vehicles and the submission of the operation plan to the supervisory government agency. Therefore, we believe that it is also necessary to confirm the scope of the responsibilities to be assumed by business operators. Next page, please.

Industry has already acquired a process to comprehensively understand that accidents do not occur by referring to international processes to ensure safety, and to show that testing is safer than humans by considering examination standards and automated driving methods. However, I think that examination and confirmation methods according to specific forms are also important for actual services.

At the same time, the specific automatic operation permission system enforced by the National Police Agency on the next page also secures safety from various perspectives, but I think it is necessary to clarify the ideal way of examination in accordance with specific cases, including the ideal way of observing traffic rules. Please go to the next page.

Finally, I will omit the details because they are repeated, but I have summarized the points to be noted from (I) to (iii). In particular, I would like to point out the second point. In the case of pursuing criminal liability for those involved, even if it is recognized that the safety of the vehicle is maintained based on the conventional method, if there is a difference in the pursuit of liability between the case where the driver drove and the case where the automated driving vehicles was used, I think it is important to clarify the scope and grounds for the difference before deepening the discussion. Regarding data sharing, the Secretariat has already presented discussion points, but regarding the scope, purpose, means, etc., the industry will cooperate toward the proposal, and regarding continuous discussion, I think the Secretariat will present an opportunity, but I hope that people who are considered to need to participate, including spectators, can participate and continue the discussion. That is all.

Mr. Kozuka: Thank you very much, Mr. . Next, I would like to move on to agenda 2 - (4). I have asked Mr. Tamamoto, Director of the Criminal Law System Office of the Criminal Affairs Bureau of the Ministry of Justice, about the current situation of the crime of negligence in the pursuit of social activities causing death or injury. Thank you very much, Mr. Tamamoto.

Tamamoto Administrator:
(Explanation based on "Material 6: Explanatory Materials of the Ministry of Justice")

I am Tamamoto, Director of Criminal Law Enforcement, Criminal Affairs Bureau, Ministry of Justice. Nice to meet you.

I would like to thank you for the opportunity to explain. I will explain the basic concept of the crime of negligence in the pursuit of social activities causing death or injury, the general flow of criminal procedures, and the current status of investigations and trials in cases of negligence in the pursuit of social activities causing death or injury.

First, I would like to explain the crime of professional negligence resulting in death or injury. Please show me page 2 of the document. When a person is killed or injured in an accident caused by a motor vehicle, there is the crime of negligent driving resulting in death or injury as a crime that can be generally applied, but this basically applies to cases where it can be assessed that a person is driving a motor vehicle. In the case of a death or injury accident caused by a automated driving without a driver, it is basically considered that the application of the crime of professional negligence resulting in death or injury under the Penal Code is problematic. In this crime of professional negligence resulting in death or injury, it is provided in the Penal Code that if a person is killed or injured by a failure to exercise due care required in the pursuit of social activities, he or she shall be punished by imprisonment with work for not more than 5 years or a fine of not more than 1,000,000 yen.

Regarding negligence under the Penal Code, it is stated on page 3 of the materials. Negligence is a violation of the duty of care, and as stated in the content of the duty of care, it is understood to be predictability and duty to predict the occurrence of consequences, and to be avoidable and duty to prevent the occurrence of consequences. Regarding predictability, it is understood in court precedents that abstract predictability is not sufficient and specific predictability is necessary. In addition, predictability is considered to be affirmed when it can be determined that the occurrence of consequences can be predicted assuming the circumstances recognized by the offender by placing a general person who corresponds to the position of the offender, such as position, age, occupation, and specialty, in the situation at the time of the act. Again, in order for a violation of the duty of care to be recognized, not only the predictability of the result but also the avoidability of the result is premised. In addition, for example, in a medical care error case, the presence or absence of a violation of the duty of care is to be determined based on the level of medical care in the practice of clinical medicine at the time of diagnosis.

Next, I would like to explain the general flow of criminal proceedings. Please refer to page 5 of the materials. When an incident or accident occurs, a police officer usually conducts an investigation and refers the incident to a public prosecutor. The public prosecutor conducts the necessary investigation together with the police officer and decides on the final disposition, such as indictment or non-indictment. If the victim is dissatisfied with the disposition of non-prosecution by the public prosecutor, he or she can file an appeal with an organization called the Committee for Inquest of Prosecution.

Next, I would like to explain the judgment of prosecution or non-prosecution by the public prosecutor. It is page 6 of the materials. In order to find guilt in a criminal trial, it is necessary to prove a crime to the extent that there is no room for reasonable doubt, and a higher level of proof is required than in a civil trial. The public prosecutor can judge whether or not the crime can be proven based on the evidence collected through the investigation, and even if it is considered that the crime can be proven, the public prosecutor can decide not to prosecute in consideration of the seriousness of the crime and the situation after the crime. After clarifying the truth of the case, the public prosecutor makes a judgment on whether or not the prosecution is appropriate by comprehensively considering various circumstances. In this way, prosecution or non-prosecution is judged on a case-by-case basis based on specific facts and evidence, so it is difficult to create a uniform standard separately for each specific case.

Next, I would like to explain the current status of the investigation and trial of cases of negligence in the pursuit of social activities resulting in death or injury. Please refer to page 9 of the materials. In cases of negligence in the pursuit of social activities resulting in death or injury, it is important to clarify the cause of the incident, the content of the duty of care, and whether or not there was a violation of the duty of care. To this end, the main investigation activities include on-the-spot inspection of the scene of the incident, collection and analysis of objective evidence such as security cameras that captured the scene of the incident, collection and detailed examination of law and general guidelines and guidelines related to the relevant work, as well as interrogation of persons who committed acts related to the incident. After these investigations are completed, the prosecutor determines whether or not there was a violation of the duty of care by the offender based on the law and evidence, and makes a judgment on whether or not the prosecution is appropriate by comprehensively considering the seriousness of the crime, the circumstances after the crime, and the like. The content of the violation of the duty of care is not specifically provided for in the Penal Code, etc., but is determined by the investigative authorities for each case based on the collected evidence and related law, etc., and whether or not the judgment is appropriate is finally determined by the court. In addition, when the prosecutor determines the content of the duty of care and whether or not there was a violation of the duty of care, the prosecutor examines the law and general guidelines, etc., and listens to the opinions of multiple experts. If there is a conflict of opinions among multiple experts, the prosecutor considers whether or not the criminal fact can be found to the extent that there is no room for reasonable doubt even based on that.

Next, I would like to explain the relationship between investigation procedures and criminal procedures. Please refer to page 10 of the materials. I am aware that in this Sub-Working Group, as written in the materials, you pointed out that if the materials and statements provided for the investigation of the accident are used for criminal trials and investigations, we may not be able to honestly cooperate with the investigation of the accident, and that the progress of the investigation of the accident may be delayed because the investigation of the accident and the investigation are conducted at the same time. I would like to explain the current situation on this point. First of all, in the investigation of the accident, I believe that various materials are submitted to the investigation agency by the people concerned, but the submitted materials are not naturally regarded as evidence in criminal trials under the current system. It is said that statements to the investigation agency cannot be used as evidence in a trial unless they satisfy the strict requirements prescribed in the Code of Criminal Procedure. In addition, the purpose and philosophy of the investigation procedure by public authorities, which is carried out to clarify the cause of the accident and prevent its recurrence, and the criminal procedure for imposing sanctions on the perpetrator are different, and it is considered that one cannot substitute for the other, and that one should not take precedence over the other. Therefore, there is no system in Japan today that places a certain limit on the investigative authority of the investigation agency in certain fields. Please turn to the next page.

As for the relationship between the accident investigation agency and the investigation agency, for example, the National Police Agency and the Japan Transport Safety Board have exchanged a memorandum as described in the documents. In this way, we have made detailed arrangements for the preservation of the accident site and the autopsy, and the fact is that the accident investigation agency and the investigation agency are conducting investigations and investigations simultaneously while coordinating with each other. With regard to the relationship between the investigation of the accident and criminal liability, in the recent Diet deliberations, in relation to the aircraft accident, there was a question about whether individual criminal immunity should be introduced in order to ensure true testimony to prevent recurrence. In response, Prime Minister responded that it is necessary to carefully consider the exemption of individuals from criminal liability in aircraft accidents, taking into account the significance and purpose of punishment as well as the feelings of the people, including the victims. In addition, if the investigative authorities were to be prevented from using materials collected by the investigation agency, there would be a Issue in relation to investigation practices, the significance and purpose of punishment, and public sentiment, including the victims. Specifically, for example, even if the accident investigation agency has established guidelines that prohibit the provision of collected materials to the investigation agency, it is considered to be of little practical benefit because the investigation agency can collect such materials through compulsory investigation under the Code of Criminal Procedure. On the other hand, if compulsory investigations by investigative authorities are prohibited by law, it will be impossible to pursue criminal liability, and the significance and purpose of punishment, such as retribution and prevention of crime, will not be realized. In addition, I believe that the question of whether the understanding of the people, including victims of accidents, can be obtained will be a Issue. Apart from serious accidents, there may be an idea that the investigative authorities will not be allowed to use the materials collected by the accident investigation agency for minor accidents. However, I think it can be pointed out that the seriousness of the accident itself will be revealed only after thorough investigation.

We will explain the systems to exempt criminal liability in fatal accidents caused by automated driving vehicles. Please refer to page 12 of the document. In this sub-working session, it was pointed out that in order to increase the incentive for development operators and companies to investigate the cause of accidents and prevent recurrence by reducing or waiving criminal liability, systems for postponing prosecution agreements and systems for exempting criminal liability should be established. First of all, I would like to explain about the current system in Japan. In Japan, as a transactional investigation method, there is a consultation and agreement system of investigation and trial cooperation type in which the public prosecutor and the suspect / defendant cooperate with each other in clarifying the facts of crimes committed by others for certain fiscal and economic crimes and drug and firearm crimes, and the public prosecutor can make an agreement on the content of the non-prosecution disposition or the request for a specific punishment for the case of the suspect / defendant in consideration of this. This system was introduced with the revision of the Code of Criminal Procedure in 2016. According to the materials submitted to the Advisory Council on Criminal Procedures Concerning the Revised Code of Criminal Procedure, which is currently held, there have been three cases since the system came into effect in 2018. When this system was introduced, there was discussion about the introduction of a self-incrimination type consultation and agreement system in which crimes that infringe on human life and body are included in the cases subject to the system, and public prosecutors are given advantageous treatment, such as non-prosecution, in exchange for admitting one's crime. Please take the next page.

As for the cases subject to the system, I have quoted the answers of the Diet, but it was discussed that it would be necessary to carefully examine whether the understanding of the people would be obtained for reducing the punishment of suspects and defendants who have committed serious crimes that protect the lives and bodies of people, such as murder, for the benefit of the law, and the cases subject to the system were limited to certain fiscal and economic crimes and drug and firearm crimes. In addition, regarding the self-incrimination type consultation and agreement system, there was a strong opinion that if the subject of consultation and agreement was whether or not one's own crime was admitted, it would lead to a situation in which it would be more advantageous to first deny the crime and negotiate with the prosecutor than to confess from the beginning, and as a result, the suspect would have to make significant concessions. As a result, the system was not adopted. With regard to the self-incrimination type consultation and agreement system, it is considered appropriate to first introduce the investigation and trial cooperation type consultation and agreement system, and then conduct consideration as necessary based on the status of its operation. Based on the above, if it is intended to establish a mechanism in which the prosecutor can reach an agreement to decide not to prosecute in exchange for the parties concerned taking measures to prevent the recurrence of accidents resulting in death or injury caused by automated driving vehicles, it is considered that there is a Issue in relation to investigation practices, the significance and purpose of punishment, and public sentiment, including the victims. Specifically, with regard to the consultation and agreement system, there is no guarantee that the public prosecutor will agree to the agreement in the first place, and in order for the public prosecutor to decide whether to agree to the agreement or to ensure the possibility of criminal prosecution in the event that the parties involved in the accident violate the agreement, it is a prerequisite for the public prosecutor to clarify the case and collect the necessary evidence. Therefore, even if a general agreement can be reached, it is inevitable that the investigation will be at least at a stage where considerable progress has been made, and therefore it is considered inevitable that the parties involved in the accident will be investigated. Based on the above, I believe that even if a consultation and agreement system is established, it will not necessarily be a mechanism that provides a significant incentive to the parties involved in the accident to work on the cause of the accident immediately after the accident. In addition, I believe that there are practical difficulties in monitoring whether the parties involved in the accident are surely implementing the agreement, such as taking measures to prevent recurrence. In addition, I have just introduced the discussion on the introduction of the current investigation and trial cooperation type consultation and agreement system, and I believe that the points pointed out there are also Issue. That's all.

Mr. Kozuka: Thank you very much, Mr. . I think it is very important to understand these systems accurately.

Let's move on to Agenda 2 - (5). I would like to hear your presentation on the fact that the introduction of the criminal immunity system does not gain the understanding of the people (especially victims). Then, Mr. Takahashi, please.

Takahashi Member:
(The following explanation is based on "Material 7: The introduction of the criminal immunity system is not understood by the people (especially victims).")

I come from a science background and have a certain level of knowledge. In addition, I have been mainly involved in the construction of systems such as the victim participation system, so I would like to talk about that point in particular. This resume is long, but to put it briefly, it is divided into Chapter 1 and Chapter 2. In Chapter 1, the conclusion that the impunity system should not be established and the reasons for it are written. There are four reasons. The first is that the understanding of crime victims cannot be obtained. The second is that it is against the legal system. The third is that it is against the right to be tried. The fourth is that it is against the normative awareness of the people, including crime victims. Next, Chapter 2 describes how to eliminate the shrinkage of manufacturers when the impunity system is not established. In the same chapter, it is written that the number of criminal liability investigations is overwhelmingly small, and that if the judgment of a specialized organization that investigates the cause and mechanism of an accident is separated from the judgment of a judicial organization, there is no need to worry about the shrinkage of manufacturers. And on the last page, some recommendations from me are written.

First of all, aside from lawyers, from the perspective of the general public and manufacturers, in criminal court procedures, I would like to talk about the fact that in the past, things that were not even expected to occur occurred. 。Until November 30, 2008, the victim had no rights in criminal court proceedings other than the right to file a complaint. Even the families of the slain were not told the date of the sentence or even the date of the trial. In addition, I did not receive a judgment, and as for criminal records, not only copying but also viewing them was not allowed as a system. In cases that received greater public attention, even the bereaved family received a numbered ticket, and if they did not win the lottery, they could not even enter the public gallery. Such pre-modern things were carried out without incident until November 30, 2008. On the other hand, since December 1 of the same year, the rights of crime victims have expanded dramatically. Of course, it has already been dissolved, but it was created by the National Association of Crime Victims, Asu no Kai. It has already been dissolved, so we will call it the former Asu no Kai from now on. The former Asu no Kai started a national campaign with 557,215 signatures aiming to acquire the rights of crime victims. At the time, there was no electronic signature and no Change.org. We collected only this amount through street signatures alone. As a result, the Basic Act on Crime Victims was established in 2004, and the rights of crime victims were born here for the first time. It was established that all crime victims have the right to have their individual dignity respected and to be guaranteed treatment appropriate to their dignity. In response to this, the first Basic Plan Review Meeting for Crime Victims began in 2005. Here, there was a discussion about recognizing various rights of crime victims in criminal court procedures. Based on the principle of the Code of Criminal Procedure, which had been in effect for more than 60 years, it was argued that the rights of the victim should be recognized from the standpoint that the victim was merely an object of investigation or evidence. Because it was such a radical revision, discussions were held a total of 11 times, and thorough discussions were held for five hours, much longer than the scheduled two hours each time. As a result, the Basic Plan for Crime Victims was approved by the Cabinet in December 2005. In this Basic Plan, it was clearly stated that criminal justice is for the rights and interests of crime victims in addition to maintaining social order. Based on this, the Legislative Council held later held discussions, and in December 2008, the victim participation system and the compensation for damage order system were established. In this way, in the past, the victim was merely an object of evidence and a subject of investigation, but after that, the rights of the victim were recognized, and the victim participation system was established based on these rights, and it has been 15 years now. If we try to establish a criminal immunity system in this context, we will not be able to obtain the understanding of the victims and the people. Regardless of whether it is intentional, gross negligence, or minor negligence, the point is that the criminal immunity system itself is not good. This is not acceptable to the people.

Next, on page 4 of the materials. In the first place, immunity from criminal prosecution undermines the order of the legal system. Because, as far as I know, there is no legal system in which immunity from criminal prosecution is provided not only in automated driving, but also in the world of medical care, the aviation industry, space development, nuclear power generation, and other fields related to science. Nevertheless, if immunity from criminal prosecution is provided only in automated driving, it will not gain the understanding of the people. In particular, in the transportation field, the number of accidents is overwhelmingly large compared to the world of space and medical care. Under these circumstances, the uniform provision of immunity from criminal prosecution will undermine the unity and fairness of the legal system.

Next, I would like to ask if this is in conflict with the right to be tried. This is my view, and in particular, what the bereaved families of traffic offenses are claiming. As I mentioned earlier, in criminal trials, the victim is not the subject of investigation. In criminal trial procedures, the victim is granted the right to participate in the trial. Nevertheless, the establishment of the criminal immunity system may deprive the victim of the right to participate in criminal trials in the first place. There is a serious suspicion that this is contrary to Article 32 of the Constitution, which provides for the right to be tried.

Finally, it goes against the normative consciousness of the people. When I listen to the discussion from a while ago and the discussion before last, I can only think that the discussion is based on the premise that the system will be completed. The problem is that in most cases, human error occurs when the system is built. If no one is responsible here, it goes against the normative consciousness of not only the victims but also the people, and it is not convincing. From the above, I believe that the elimination of atrophy for the development of science and technology should not be done by establishing a criminal immunity system, but should be done by the method described in Chapter 2, that is, by creating a specialized accident investigation committee.

I will explain page 6 of Chapter 2 of the document. First of all, I would like to state the premise that the number of cases in which criminal liability was pursued in relation to science and technology is actually overwhelmingly small. It is no exaggeration to say that there are almost none. Nevertheless, I can only think that the reason why manufacturers are shrinking is because they fear the loss of business due to the huge amount of recall costs, followed by the filing of a shareholder representative lawsuit. Of course, there are many civil lawsuits involving medical care errors, but in criminal cases, the Public Prosecutor's Office does not prosecute unless there is evidence that leaves no room for reasonable doubt from the viewpoint of the modesty of the Penal Code. I would like you to understand that point. I would like you to see an actual case. In the medical care incident, the incident of erroneous intravenous drip of the Tokyo Metropolitan Hiroo Hospital can be mentioned. It is an incident in which a disinfectant solution was administered intravenously when a normal saline solution should have been administered intravenously. At Yokohama City University, it was a case that they operated on the wrong patient. There have been cases in which a tracheal intubation tube was inserted by mistake. In this way, in reality, only cases that are too shoddy are prosecuted. In addition, there was a fraud case in which it was good to treat the teeth to correct atopic dermatitis. This was also booked as a criminal. In this way, in most cases, criminal liability is pursued in the surrounding areas, not in the scientific findings themselves. I would like you to understand that scientific knowledge itself is not disputed.

Next, on page 8 of the materials, on the other hand, it is true that there are cases in which scientific knowledge has been disputed. Typical examples are the Mitsubishi Fuso Truck Bus Incident and the Kyorin University Hospital Waribashi Incident. I think you all know these incidents. I would like to make a conclusion first. In the world of science, I believe that a separate organization such as a third party committee or an independent administrative committee should be established to investigate the cause and mechanism of the accident, and that the judicial body should respect the judgment on the science. Based on this premise, I believe that negligence, causality, predictability as stated by the Ministry of Justice, possibility of avoiding consequences, violation of the obligation to avoid consequences, and causality should be subject to the exclusive jurisdiction of the judiciary. However, although the judicial body should respect the judgment of the third party committee and the independent administrative committee on the investigation of the cause and mechanism of the incident, I do not think it is possible to recognize the legal binding force because it may violate Article 76, Paragraph 2 of the Constitution, which states that public authorities cannot conduct a trial as a final instance. I will tell you a specific case. It is the Mitsubishi Motors Recall Concealment Incident on page 9 of the document. This is an incident in which the part of a large trailer that links the wheels and axles, called the hub, was damaged, the left front wheel came off, and a housewife who was walking was killed and children were slightly injured. Before this incident, there had been many similar incidents. We did not recall all of these and hid them within the company. However, a fatal incident actually occurred, and it became a social problem and became a major treatment. The manufacturer's claim was that the damage to the hub was wear and tear caused by poor maintenance by the user. In response, the prosecution argued that the hub was not strong enough at the manufacturing and design stages. The District Court, the High Court and the Supreme Court have all found him guilty. There was no doubt that the hub was damaged by metal-fatigue, but the point of contention was whether the metal-fatigue was caused by poor maintenance by the user or a defect in the original design manufacturing process. As the reason for the Supreme Court's finding of guilt, it is emphasized that in the first place, the hub is called something that lasts a lifetime in the sense that it will never be damaged until the vehicle is disposed of. Nevertheless, there were actually 16, 18, or more hub failures before the accident occurred. From this, it can be inferred that it is a fundamental defect at the design manufacturing stage. In fact, the manufacturer did not implement a test method for strength durability. As was the case in medical care, the Supreme Court does not accept the logic that the causal relation cannot be proven because there was a shortage of materials and a shortage of materials due to a failure to perform the prerequisite test or duty of care. The fundamental idea of the Supreme Court is that if there is a shortage of materials, the person who caused the shortage of materials should be held accountable. For these reasons, the Supreme Court recognized that the cause of the damage to the hub was a fundamental lack of strength at the design manufacturing stage because accidents caused by the hub had occurred many times despite the fact that it was originally called "one life" and because the manufacturer had not conducted a strength and durability test. Therefore, in order to prevent this from happening, we should have taken improvement measures such as a recall, and we have imposed criminal liability on him for failing to do so, which is a violation of his duty of care. Next is page 11 of the document. In this regard, it should be noted that this incident is not necessarily a case of willful misconduct or gross negligence. The concealment of the recall is certainly intentional. However, it was not intentional to cause the death of someone. If it is willful to cause the death of a person, it is murder. This case is professional negligence resulting in death. Professional negligence resulting in death is different from gross negligence. Therefore, I would also like you to understand that this is not necessarily a case of willful misconduct or gross negligence. In other words, even if it is a case of ordinary negligence, if the criminal exemption system is established uniformly, there is a fear that even such a case as the Mitsubishi recall accident will be exempted from prosecution. This will not gain the understanding of the people.

Next, on page 12 of the materials. Here, I am writing that as one method, plea bargaining is possible, but it is not impossible. However, as the Ministry of Justice said, if this plea bargaining is a mechanism in which people tell the truth in order to exempt themselves from criminal responsibility, it will not be possible to obtain the understanding of the people. There should be an administrative manager and a director in charge of Technology development, so if those people are truly responsible, in order to punish such a huge evil, in exchange for confessing the fact of the crime of the huge evil, I think it is possible to make a deal to obtain criminal immunity for themselves, or to receive a non-prosecution disposition or a lighter sentence. Even in that case, it will not be possible to obtain the understanding of the people simply by punishing organization and corporations. After all, I think that it is necessary to establish a dual punishment rule that punishes both individuals and corporations, who are administrative managers behind the huge evil. Now, the term organization punishment has been mentioned in various opinions, but I would like you to understand that this is a dual punishment rule, not only punishing organization.

Next, regarding the Kyorin University Waribashi case, I don't intend to pick a quarrel with the Public Prosecutor's Office, but to be honest, I feel that I have often prosecuted such a case. The prosecution's argument was that the disposable chopsticks stuck in his throat, reached his brain, bled into his cerebellum, formed blood clots in his cerebellum, pressed on his brain stem, stopped breathing, and died. On the other hand, as shown on page 14 of the materials, in fact, not only the cerebellum but also a large vein called the cerebral sinus, which flows into the brain, was pierced by a disposable chopsticks, and it was blocked. Blood flows from the heart to the head, but the blood goes out of the head through the veins. Because the vein is blocked, all the blood stays in the head. In addition, fluid in the blood plasma collects in the brain, causing edema. Then the brain swells up. As it is written on page 15 of the document, it was clear at the time of the forensic autopsy that the average weight of a child's brain was 1240 grams in the first place, but when the forensic autopsy was performed, it was 1510 grams. On the other hand, the blood clots in the cerebellum weighed no more than 26 grams. Just looking at this, I think at the time of the judicial autopsy, the prosecutor had to give up the mechanism of death due to cerebellar blood. And in the case of a cerebellar hemorrhage, the cerebellum is on the outer surface of the back of the head, so it can be easily removed by surgery. However, in the case of the cerebral sinus, this is deep, so it is very difficult to remove the thrombus here. There is little chance of saving lives. Even though it was clear from the autopsy stage that there was no possibility of saving his life, he prosecuted him. Therefore, I think that the prosecution should have heard from various doctors in advance and humbly listened to them about their medical knowledge. Had I heard this, I don't think I would have been prosecuted.

Based on the above, I would like to make a final recommendation. On page 16 of the handout, we should set up a specialized accident investigation committee. And when it comes to the causes and mechanisms determined there, I think the judiciary should respect them. This is the domain of science, so lawyers should not enter it easily. Otherwise, you will be falsely accused. However, since there is Article 76, Paragraph 2 of the Constitution, it should not be legally binding. On the other hand, predictability, possibility of avoiding the consequences, violation of the obligation to avoid the consequences, and existence of a causal relationship are matters that should be judged by a lawyer because they are laws. And the important thing is what to do with the members of the committee. I think 9 or more members are good. This is because the Aircraft Accident Investigation Commission has nine members. And in fact, there are 9 people in the citizen judge trial in our country. The lay judge system in France is also 10 persons. In other words, it is often 9 to 11 people. So, I think that number of people would be good. And it should be unanimous. Because by majority, what is not true does not become scientific truth. If scientific opinions do not agree, it should be written both ways. In addition, it is important to leave the materials for other scientists to validation. In fact, most of the actual brain specimens were destroyed in the Waribashi incident. In this way, I cannot make a validation later. We will make sure to keep the materials and disclose them so that third parties can make a validation later. Finally, I would like to say to the Ministry of Justice that there are many specialized departments in the court, such as the medical care Concentration Center and Specialized Department, the Intellectual Property Department, the Labor Department, and the Administrative Department. The Metropolitan Police Department also has a special department called Cyber Unit. Nevertheless, only the Ministry of Justice and the Public Prosecutor's Office do not have a special department. It is true that the Public Prosecutor's Office also has a Criminal Affairs Department, a Special Investigation Department, a Traffic Department, and a Trial Department. However, I am sorry to say this, but in the case of the Deputy Prosecutor of the Traffic Department, I think that the scientific knowledge possessed is insufficient. I don't think it has reached the level of the Accident Investigation Commission. Therefore, the Public Prosecutor's Office should also establish a scientific investigation department specializing in science. By doing so, unjust prosecution can be avoided, and I believe that the understanding of the people, particularly the victims of crime, can be obtained. That's all.

Mr. Kozuka: Thank you very much, Mr. . I also had many things to think about. I received 4 presentations today, and I learned a lot from them. In particular, there have been various discussions on impunity since the first edition, and I was concerned that various people might be using this word in various ways. Since I am also a lawyer, I believe that immunity from liability refers to the exemption from liability in cases where liability would have arisen based on the law at the time of the accident, or the creation of a special law concerning automated driving to prevent liability from arising only in relation to automated driving, where liability would have arisen under basic laws such as the Civil Law and the Products Liability Law, whether in criminal or civil cases. However, in the discussion, it was confirmed that there would be no criminal liability in such cases, and in the process of creating a system, there were opinions that it would be better to have a system in which there would be no liability in such cases. These are stories where liability does not occur as a system design rather than immunity. I felt that it would be better to use these words differently. In addition to the above two, as explained today by the Ministry of Justice, there are responsibilities that should be pursued in the course of the procedures. However, if certain conditions are met, in particular in the case of a criminal, there is a disposition of non-prosecution or suspension of prosecution, which means that the criminal is not prosecuted. In that sense, immunity is sometimes called. Including this point, I think it would be better to separate the terms properly. In particular, it seems that the sub-working group is drawing attention from society, and I think that the general public will also see the minutes, so it is my impression that it is better to clarify what kind of discussion is being held, not necessarily only legal and technical discussions. I am not sure if this is related to this, but there are some members who have stated that they would be exempt from liability at previous meetings, but have requested that they would like to add a little more to the purport of this. I believe that the materials have been prepared and distributed to all of you in advance. First of all, I understand that you would like to clarify your intention of immunity, as you have been saying, by having Ochiai members clearly state their intention based on the materials. Could you please make some comments?

Ochiai Member: . Now, I would like to make a brief statement. First of all, regarding the discussions at this meeting, at the first meeting, I stated both the points concerning the future and the points related to this discussion. Among them, I think that the goals of this discussion were a little unclear. In addition, while participating in the discussions afterwards, I feel again that there are quite a few cases where it is difficult to reach a conclusion. Among them, particularly for detectives, I think that we should sort out the content of the errors and what kind of duty of care they have. In that sense, I think that we are at the stage of further narrowing down the content while conducting social implementation at a certain reasonable time and place. Please show me the next slide.

After all, it has already been discussed, but as the Japan Automobile Manufacturers Association explained today, it is important to determine where the target point is in Technical development as a prerequisite social condition. It is important to develop guidance, etc., and we are discussing about negligence, but as a result, I believe that the Ministry of Land, Infrastructure, Transport and Tourism will approve these matters, etc., so I believe that there are matters that should be examined in accordance with where the level is at that time. I believe that basic matters can be what kind of performances are actually required, what roles are required of development users and operators, what kind of actions should be expected of other road users, and what to think about when information input or assistance is received from roads, etc. If such matters are not determined, I believe that what should be done will not be determined. Please show me the next slide.

In particular, regarding the guidance and approval of the public authorities side, I am aware that there are parts that cannot be done because there are normative elements in the end, but I believe that it is desirable to quantify the parts that can be done. In fact, I talked about the National Strategic Special Zones, but I think that environmental improvement for rational social implementation can be considered, such as setting a place where people are not expected to enter, in the sense of expectations in the world, such as implementing it by dividing the place by highways, etc. Please show me the next slide.

This is a point for the future, and as is the case this time and in the future, I believe it is important for the development side to be aware of incentives to make improvements to the maximum extent possible and to make institutional design. Incentives are written as incentives, including the fact that we will have to work on them by strengthening penalties. In that context, I believe that creating accident investigations and information linkages systems as discussed at the previous meeting will lead to the information linkages of close calls, appropriate clarification of the actual situation at the time of actual accidents, and appropriate punishment. In the future, I believe it is necessary to firmly establish a strong sanction system for corporations and an obligation to cooperate in accident investigations.

Mr. Kozuka: Thank you very much, Mr. . In relation to the arrangement I thought earlier, is it correct to understand that it is better to have rules that clarify the parts that are not responsible?

Ochiai Member: Yes. In the current situation, I think it is difficult to immediately revise the law, so I think it would be good to clarify where the duty of care lies, and to discuss in the direction that Chief Kozuka said.

Mr. Kozuka: Thank you very much, Mr. , a member of the Advisory Committee, also stated that he would like to supplement his statement of immunity. Could you please do so?

Mr. Suda: Today, I learned a lot from your very valuable opinions. Thank you very much. I think I spoke from two perspectives. One is that we are investigating the accident, and there is a risk that we will not be able to obtain accurate testimony, so I made a statement in the sense of plea bargaining. The other is from the development side, and as the chairman said, we will create a rule that says you are not responsible. The purpose of the statement was to clarify that if you follow the rule and obtain design and approval, you are not responsible if you meet the conditions. Of course, I do not believe that malicious or intentional cases should be excused.

Mr. Kozuka: Thank you very much, Mr. , please speak in this order.

Takahashi Member: It is true that the discussion and definition of immunity from prosecution are complicated, and I think it must be sorted out. Listening to the manufacturer, it is true that if the system program is done properly, there are almost no accidents. However, that is the conclusion, and the problem is whether the system can be done properly and whether the algorithm can be done properly. The actual mistakes occur when you put in the wrong equation, put in the wrong number, put in the wrong unit, do something wrong, and so on. Therefore, we must make efforts to avoid mistakes at the program construction stage. For that purpose, it is meaningless to examine civil court cases as much as possible. In civil cases, the burden of proof is shifted by Article 3 of the Self-Compensation Act, so it is de facto starting from the fact that there is a cause of liability. In other words, the main issue of civil suits is the amount of damage. Therefore, in order to determine in what cases accidents occurred and who was negligent, I think it is meaningless unless we collect criminal court cases. This is the first point. validation

Second, regarding the scope of civil liability, under Article 3 of the Self-Compensation Act, the operator is the responsible entity. If it becomes a automated driving, I think it is fine to replace the term "operator" in Article 3 of the Self-Compensation Act with the development operator of the manufacturer, the technology development operator, or the manufacturer itself. If so, I think there will be no contradiction. In addition, there is an opinion that it may be fine to guarantee by a government guarantee project, etc., but the current government guarantee project is insufficient. The maximum amount of the government guarantee project is only 30 million yen. Moreover, unlike compulsory automobile liability insurance, it also provides negligence offset. In this regard, compulsory automobile liability insurance almost does not provide negligence offset. And in the case of voluntary insurance, the amount of damage is about 60 million yen to 100 million yen in any case. Therefore, if you really want to guarantee the victim, I think it will not be helpful to relieve the victim unless you create a further expansion of the government guarantee project.

Mr. Kozuka: Thank you very much, Mr. , Mr. Inadani, please.

Inadani Member: Thank you very much for your various reports today. I learned a lot. Later on, I would like to reiterate my comments on the report from the Ministry of Justice. I have never used the term "immunity from prosecution" in the sub-working group. However, I would like to clarify the scope and justifiable cause for noncompliance with the law of liability. First of all, I have never made a comment on personal immunity from prosecution. I made a comment to the effect that it is necessary to clarify the scope and grounds of liability. In addition, I have been saying that corporate sanctions should be strengthened. However, criminal sanctions have a strong aspect of social condemnation, and it is quite difficult to use them functionally. Therefore, from the perspective of utilizing administrative sanctions, I would like to clarify here that I have proposed the method of deferred prosecution agreement as a method to improve security as a social system. In addition, I will comment again later that the deferred prosecution agreement is a system that is quite different from mere immunity from prosecution. Before that, I would like to comment on all of the presentations.

First of all, I agree with the preliminary and general evaluation approach announced by Mr. Fujita. In light of the current development methods and performance management methods of AI, it has been pointed out that if AI is made to behave in the same manner as humans, performance will inevitably decline. Therefore, based on an appropriate cost-benefit analysis, as Dr. Ochiai mentioned earlier, it is desirable to develop safety standards for programs in a quantified form, and to take a proactive and general evaluation approach that determines that a program is defective if it does not meet the safety standards. In that case, I think you pointed out that there may be some parts that are not guaranteed. As Takahashi members said earlier, I think it will be a problem to consider a generous compensation measure. On the other hand, in relation to product liability, even if various simulations and demonstration experiments are performed before the product is released to the market and the "defect" is eliminated by the prior and general evaluation approach at that stage, I think that actual driving behavior may appear in a form that is not predicted or intended, and as a result, the product may be "defective." In that case, I think the topic was how to understand the time when the defect occurred in relation to the delivery, but I think that point will be a problem. I believe that it is possible to consider that there will always be "defects" without fixing them at the time of delivery. However, I believe that there are certain parts where we have to consider the defense of development risk to a certain extent. Based on this issue, I believe that it is necessary to further clarify the scope of responsibility and discuss how to consider public compensation measures in such cases. In addition, in the case of a methodology in which "defects" are defined and the scope of responsibility is determined using a prior and general evaluation approach in a form linked to safety standards, I thought it would be good to include a mechanism that can update the safety standards according to changes in the situation of the technical development as a set. I wonder if we can avoid the obsolescence of regulation by doing so. To that end, as the Secretariat announced today, I believe it is necessary to carefully implement systems for the provision of information and design incentives for that purpose. As it was the last slide, there may be cases where it is difficult to prove the burden of causality or the defect itself. In this case, we thought it would be good to create a system in which the results of the expert judgment made by the accident investigation agency are reflected in the fact finding in some form.

As for the story of the Japan Automobile Manufacturers Association, in the matters to be confirmed on page 6, I think what is different from the current law was the problem. In the case of the current law, I think that the story called risk transfer functions as a responsibility demarcation point. In the case of a car that has satisfied the safety performance and has been delivered, the risk of the car has been reduced to a level that can be managed by the operator, so the driver will be responsible for managing the risk after the car has been delivered. However, regarding the unmanned automated driving, as Mr. Fujita explained in relation to the interpretation of the Automobile Liability Act, the operator basically cannot control the danger, and the danger existing at the stage of the development operator is expressed as it is. In particular, it is premised that the operator will not cause any problem, but as a result, the management of the risk remaining after distribution will be directly asked to the development operator, so I think the way of sorting out the responsibility is a little different from the past. However, if we use the concept of defects in the sense that Mr. Fujita was organized based on the prior and general evaluation approach, we may be able to say that there is no change in the scope of responsibility in the abstract in the sense that we are not responsible if there is no defect. On the other hand, as you pointed out, in the four quadrants, for example, in the case of sudden jumping out, I think there are cases where the judgment on who should manage the risk can be divided depending on whether the person who jumped out is a child or an adult. Therefore, I think it is necessary to clarify who should manage the risk and in what form, including discussion on how to improve the driving environment.

Here are my comments to the Ministry of Justice. The chairman pointed out that checking the facts is one important point, so I will share the facts as far as I understand. As stated on page 6 of the Ministry of Justice document, it is absolutely true that the decision to prosecute or not prosecute is made in accordance with individual circumstances, and I have no objection to that. On the other hand, I believe that even in the United States and the United Kingdom, where prosecutors have a wide range of discretion to prosecute based on the principle of party to the case and the principle of prosecution expediency, there are actually detailed guidelines for the discretion to prosecute depending on the type of crime. In particular, in cases where a high level of expertise is required, it is my understanding that there is a strong tendency in comparative law to develop prosecution discretion guidelines according to types. In that case, when determining the scope of criminal responsibility in the context of unmanned driving, I also mentioned one justifiable cause for noncompliance with the law and said that it would be good to do so. It cannot be denied that further examination is necessary in criminal law theory on the issue of whether or not it is determined by the presence or absence of justifiable cause for noncompliance with the law. However, at least the circumstances listed as matters to be considered seem to be consistent among various criminal law scholars, so I believe that it is possible to sort out to some extent what kind of circumstances affect the judgment of prosecution in the context of unmanned driving. At that time, in relation to the points pointed out by the members of Ochiai, if the determination of the presence or absence of defects is linked to the formulation of safety standards for programs, I think it is important to link it to the formulation of guidelines for development operators and operators in a manner that is consistent with these standards. As suggested by Mr. Takahashi, if we were to develop a system such as pre-investigation in serious cases in relation to pre-investigation and the utilization of the results of the investigation, I thought that considering the prosecution authority of the specialized agency that conducted the prior investigation as in Article 96 (1) of the Antimonopoly Act could be considered as a method in relation to the prosecution discretion. Next, on page 12, the postponement of prosecution agreement system and the criminal immunity system are written side by side, but I believe that the two systems are quite different in nature, and that they are also different in nature from the current consultation and agreement system, so I would like to add a few remarks on this point. There are various forms and variations of the deferred prosecution agreement system. In addition to the voluntary provision of information by the company prior to the investigation by the authorities, the deferred prosecution agreement system is a system in which the company is required to implement a considerable variety of problem-solving measures as necessary in light of the nature of the case, including the implementation of the reorganization of the organization system, etc., as raised by the Takahashi members, the payment of compensation to the victims and fines to the government, and the appointment of a supervisor to monitor the status of the fulfillment of the promise. Therefore, in a malicious case, the burden on the company is considerable, and it is basically understood as a type of effective sanctions. I think it is better to understand that it is a mechanism to promise that you will do everything necessary to solve the problem and make sure that you will do it, although you will not prosecute, rather than just talking about immunity or trading to exempt you. I would like to repeat what I said at the beginning. Criminal punishment has the meaning of social condemnation. As you pointed out today, I believe that there are points where we should be cautious about using criminal punishment from a functional perspective in consideration of public sentiment. Therefore, I think it is good to arrange the Agreement on Postponement of Prosecution as a system of administrative sanctions against corporations, and the substantive constituent elements thereof should be based on the objective safety of unmanned vehicles as proposed by Mr. Fujita, and based on the fact that it is an administrative sanction against corporations, it should be arranged well so as not to be intentional or negligent, and it should be utilized well.

Finally, to the Takahashi members. I think that the mechanism in which the accident investigation is preceded as you proposed is very advantageous in cases where high expertise is required for scientific knowledge, such as unmanned operation. In particular, with regard to unmanned operation, there is a possibility that overseas companies will come in, and there is a possibility that a global investigation will be required. In addition, how to collect data inside a huge company, or if there is actually a malicious individual who is active in it, information on that individual will be collected. As stated in the materials by Yoshikai, I think that the provision of information by companies will play an important role. On the other hand, as stated in Mr. Suda's previous statements, I believe that those who are active in the investigation field actually feel that there are parts where cooperation with the accident investigation and incentives with other systems are not being taken well. Regarding the issue of incentive consistency, I believe it is important to ensure consistency for companies rather than individuals. Again, companies have the most materials, so from the perspective of adjusting incentives, I think it will be a very important solution to successfully introduce the system of deferred prosecution agreement as I mentioned earlier. In my opinion, the system of deferred prosecution agreement should be utilized in a way that contributes to both the clarification of facts and the realization of justice. If the system of deferred prosecution agreement is introduced in this way, it will lead to the clarification of the truth of the case, and information on malicious individuals will be properly obtained. Therefore, from the viewpoint of realizing appropriate criminal punishment for individuals, the system should function as a system that ensures the cooperation of companies that have a large amount of information. In addition, it should be effective from the viewpoint of adjusting incentives when an accident investigation is preceded and from the viewpoint of realizing effective criminal punishment.

Mr. Kozuka: Thank you very much, Mr. . Then, I would like to start with Mr. Sato.

Member: Thank you very much for your valuable presentation today. First of all, in terms of your impression or comment on the content, I believe that a prior approach is better for automated driving. With regard to the opinion on the point you pointed out, as a premise of the idea, even if we change the way of thinking about defects, it is not the approach of expanding the number of self-indemnification operators that Takahashi members said, but as arranged by the Ministry of Land, Infrastructure, Transport and Tourism, we understand that it is fine to discuss the requirements for defects under the Product Liability Act after responding to the current self-indemnification even at Level 4. In addition, as a problem of the ex post facto approach, you pointed out that the operator will be forced to guarantee the content of the vehicle control in a state where it can be relied on by the program in the first place. I understand that the Ministry of Land, Infrastructure, Transport and Tourism's arrangement allows the operator to be primarily responsible, and from that perspective, I don't think it is that much of a problem. As for the prior approach, in the end, I believe that those that are guaranteed to be safe through the prior approach will be liable under the Self-Compensation Act, so I understand that this itself will not be such a big problem. If so, it will be a matter of claiming against the automobile manufacturer, etc. after the insurance payment by the insurance company, etc., and I thought it would be different from the viewpoint of proof in that aspect, but I don't think it is necessarily clear that the prior approach makes it easier to prove the existence of the defect. Next, regarding the ex post facto approach, I believe that it is a difficult task to determine the criteria for standardizing what kind of programs are quantitatively safe and whether they are significantly safe compared to humans. Among the problems you pointed out, if the driver is a human, he or she will be responsible for the accident. However, if it is possible for a human to avoid the accident, I believe that the accident is foreseeable and avoidable. If so, there is a possibility that it will be said that it will be an object that should be prevented in advance by simulation, etc. If it is appropriate to allow only the four quadrants that are difficult to predict and avoid, which is also in the materials of the Japan Automobile Manufacturers Association, I thought that it is necessary to crush the predictable parts in advance through a simulation that is usually conducted while taking a preliminary approach. Regarding the formulation of scenarios, as pointed out by the Japan Automobile Manufacturers Association, I think it is a difficult issue to what extent we should consider the behavior of other traffic participants, for example, and I can understand that if it is a sudden rush, it is unavoidable even if it cannot be avoided. On the other hand, with regard to the extent to which the illegal acts of other traffic participants should be taken into consideration, for example, it is reasonably possible that there are people who are crossing at a speed while looking left and right at a place where crossing is prohibited, or there are people who are sleeping on the road late at night, so I think it is difficult to consider how much should be considered to be predictable and avoidable by looking at past court cases. Although it is a little off topic, I think it is also a problem how to think in relation to some abstract requirements in the safety standards when an accident occurs after all even if various scenarios are combined. For example, while there are somewhat abstract requirements such as "when operating, it conforms to the provisions of the Road Traffic Act" and "there is no risk of interfering with the safety of other traffic," I think it is very difficult to determine to what extent the quantitative safety that we have just discussed will be consistent. Regarding the abstract safety standards mentioned earlier, if an accident occurs as a result, looking back later, there is a possibility that the accident does not satisfy the safety standards formally, for example, it may be a formal violation of some regulations of the Road Traffic Act, and then it may be recognized that there was a defect. In this way, as long as the current security standards are vaguely defined, I believe that it is necessary to create certain guidelines, including the scenarios I mentioned earlier, and discuss the level and sense of level at which defects do not occur. Along with this, I believe that we should discuss, for example, if it is possible to create guidelines and the like, to what extent negligence and various requirements under the Penal Code should be violated, rather than talking about impunity.

Mr. Kozuka: Thank you very much, Mr. . Then, Mr. Yoshikai, please come in.

Yoshikai Member: I would like to make a supplementary note to the comments I made this time. As stated earlier by the Chairman, I believe that we should distinguish between cases in which liability is not pursued even if a person is found guilty, and cases in which liability cannot be pursued in the first place. However, if liability can be pursued but immunity is granted, it is possible to consider cases in which the injury or injury is relatively minor, and in such cases, there is a provision that the person is exempt from punishment for driving a car, so I think this is helpful.

Second, regarding the relationship between the opinions of experts and criminal liability, Mr. Hatano of the Japan Automobile Manufacturers Association (JAMA) also made various reports, and I thought that there would be various concerns from the manufacturers. However, criminal trials and criminal prosecutions are both based on facts, so one of the facts is that the opinions of technical experts are very important, and it can be said that ignoring the opinions of technical experts and proceeding forward is rather an exceptional case. As reported by Mr. Takahashi, it is natural that the fact that criminal liability can be pursued means that the case is a poor case or a case that is quite far from the general standard. At such a time, if it can be argued that the case is not far from the general standard, I believe that there is no need to shrink from pursuing criminal liability.

Third, in relation to plea bargaining, I believe that it is rather important to collect objective evidence. Mr. Inadani has just mentioned the postponement of prosecution, but the postponement of prosecution is premised on the fact that prosecution can be carried out after all, and it is not possible to postpone prosecution if prosecution cannot be carried out. There was talk of postponement of prosecution as an incentive for investigation, but rather, I believe that the original form of the manufacturer is to make it impossible to prosecute in the first place by answering to the investigation committee that it had made a legitimate design, produced a legitimate product, and used a legitimate product when the accident occurred, as mentioned by Mr. Hatano earlier, and that cooperating in such a way to make it clear that it is not responsible can be an incentive.

Mr. Kozuka: Thank you very much, Mr. . It's already time, but could you allow me to extend the time by 10 to 15 minutes? Mr. Goto, please.

Member: Thank you for your various reports today. As Chairman Kozuka summarized in the middle of the meeting, I have long felt that the meaning of the term immunity from prosecution should be used more carefully. This time, Mr. Yoshikai told us whether the crime of professional negligence resulting in death is established in the first place or there is negligence, and whether a plea bargain is made to obtain information or exemption is made in the case of minor damage. I believe that it was a major achievement this time that I was able to clearly distinguish and discuss them.

I think all of these points are important, but based on what I heard from the Japan Automobile Manufacturers Association today, I believe that the first concern of the Japan Automobile Manufacturers Association is whether or not involuntary manslaughter in the pursuit of social activities will be established. In that regard, considering the story of predictability and the possibility of avoiding the result, we understand that it will not be established if the result is unavoidable in the first place, and if it is difficult to predict in the first place, I believe that involuntary manslaughter in the pursuit of social activities will not be established in general. In that case, I think the remaining part will be the upper left quadrant of the document. I think that manufacturers should respond firmly to this area, but Dr. Takahashi pointed out that there are cases where it does not go as expected.

Regarding the derailment incident of the Mitsubishi Fuso Truck that appeared in Mr. Takahashi's talk, as you pointed out, it was a concealment of the recall, and I do not think there was any predictability in terms of who would be hit and lost. However, I understand that it was recognized because there was sufficient predictability in a broad sense that if it was left as it was, serious accidents would eventually occur. I believe that the Japan Automobile Manufacturers Association is also sorting out the matter as something that should not have happened in the first place, but I understand from today's announcement that if a similar incident had occurred, it would be unavoidable even if a crime of professional negligence resulting in death was established. I believe that there is no difference in recognition on the point that there should be no such thing as a concealment of the recall, and if there is, of course, criminal punishment will follow. In the case of automated driving vehicles, if an incident occurred and it was revealed that there was actually a problem with the program, it may be possible to respond by online software update without recalling, and if it was not done in a timely manner, it is quite possible that it would be a business mistake. I believe that the Japan Automobile Manufacturers Association will probably be convinced on that point as well.

The problem is that there are hidden errors that were not noticed at first, and accidents occur due to them, and the errors are discovered by chance. There is a question as to whether it is a business mistake from the beginning to overlook the errors. I think it is related to Mr. Fujita's talk about the definition of defects, but from the perspective of a normal car that is not a automated driving, it probably cannot be said to be a business mistake. Of course, it is premised that administrative safety standards and the like are met, and that the manufacturer takes inspection processes to prevent such errors from occurring, but my sense is that accidents that inevitably occur are not called errors at the beginning. If so, there are few problems, and we are talking about how to prevent the second and subsequent accidents, but even so, there are no clearly confirmed events, which is the indication from the Japan Automobile Manufacturers Association and Mr. Ochiai, and I think it is the source of anxiety. There is no need for legislation on this, and it may be a restatement of previous precedents, but I think it is sufficient to confirm it.

In such a case, if criminal liability is not sought for the accidents that are foreseeable and avoidable but inevitably occur, it will be a question of how to secure civil liability and administrative liability. For example, in the sense of making it easier to pursue product liability, it may be possible to shift the burden of proof of defects to the manufacturer. Mr. Fujita pointed out that it is strange that drivers or automated driving vehicles users are responsible for them at that time. If the manufacturer can promise to respond firmly, I think we can do a little more than the current government guarantee.

In summary, I have the impression that there are not so many points of conflict. I think the biggest point of conflict is whether or not to carry out strict criminal immunity and plea bargaining in order to collect data in criminal proceedings. However, this will become a major problem in Japan's criminal and judicial systems, so I think it is necessary to discuss this separately from the story I just mentioned. That is all.

Mr. Kozuka: Thank you very much, Mr. . Mr. Yokota of the General Insurance Association of Japan, could you please make a statement?

Member: Thank you very much. I am Yokota of the Non-Life Insurance Association of Japan. Thank you very much to everyone who made a presentation. I would like to briefly refer to Mr. Fujita's materials from the perspective that the Self-Compensation Act and compulsory automobile liability insurance are mentioned, or that the non-life insurance industry usually responds to accidents.

Among the two approaches, I have some concerns about the points described in the Problems with the Ex-ante and General Assessment Approach. I believe it is essential to take measures to ensure prompt relief for victims, equivalent to the current measures, so that relief for victims will not be reduced compared to the case of manual driving. One is that it is impossible to prevent accidents that can be prevented by human drivers with an extremely low probability. I believe that the establishment of evaluation criteria is a probabilistic evaluation, but depending on the setting, it is possible that relief for victims will not be provided, so I believe that the establishment of evaluation criteria is extremely important. In addition, regarding the Government Security Project, I would like to comment from a perspective other than what Dr. Takahashi said earlier. In the case of the Government Security Project, only victims can claim for the project, and after the victims pay for the treatment and other expenses, they will claim for the Government Security Project at a later date, so there is a Issue that prompt relief for victims may not be provided. I believe that it is necessary to consider this while taking into account the Issue.

Regarding the problem of the ex post facto and individual evaluation approach, it was mentioned that the content of vehicle control by the program should be guaranteed in the case where the operator is an individual. However, if the part of data utilization, which was the subject of the previous discussion, is organized, I think that the person who should be responsible will be more clarified, and I think that it will be reduced to have an individual assume the responsibility. That is all.

Mr. Kozuka: Thank you very much, Mr. . Next, Mr. Nishinari, please.

Nishinari Member: Thank you, . I would like to make two brief comments.
First of all, the four quadrants of the JAMA materials were easy to understand. However, I think it is most difficult to enter into any quadrant. For example, cyber attacks are skillful and can be made to look like a machine failure. As was a topic of conversation the week before last, when a special electromagnetic wave is applied to the lidar sensor, which is the eye of automated driving, an object can be erased, and it is possible to make it run as if there was nothing there. I think it is necessary to collect data thoroughly in case such new technology emerges.

Another point has not been discussed today, and I am not sure if it will be discussed in the future, but I believe that prevention is the most important. Therefore, I think that an organization that collects close call data is also necessary. It may not be related to the talk about criminal immunity from prosecution earlier, but I thought that a mechanism that anonymously collects data on events that do not result in an accident but are judged to be dangerous and shares it with manufacturers and others is also necessary. That is all.

Mr. Kozuka: Thank you very much, Mr. . Although I extended my time, I think we were able to have a productive discussion today.

I organized the meaning of the term "immunity" in the middle of the meeting. In addition, Dr. Nishinari mentioned extreme examples such as cyber attacks. I believe that development sites, including automobile manufacturers, are concerned about how much cyber attacks, what kind of situations they need to prepare for, and how much they need to respond. I believe that they are concerned about the limit of liability in terms of how much they should respond to avoid being negligent. I am not sure if this will eventually become a guideline, but I feel that there are parts where it is difficult to move forward without a certain degree of consensus. On the other hand, in the process of fulfilling these procedures, if there are cases in which experiments that should have been conducted have not been conducted, or if false data have been produced, I do not believe that anyone would deny that there is liability, including criminal and civil liability. I understand that Mr. Takahashi has repeatedly said that if it was done under pressure such as putting priority on profit in the corporate constitution, it would be the responsibility of the entire company and could be the responsibility of the upper management. I felt that there was no one who would deny this point. After clearing all those problems, I think there is the problem that Mr. Fujita mentioned at the beginning today. For example, there is a problem of what kind of program should be prepared for the case where a child runs out and cannot stop because it is unmanned, which a human driver can notice and step on the brake. As an extreme solution, for example, it can be considered that the vehicle is equipped with a switch that can be pressed by a person in the vehicle, not limited to the driver. However, if you install it, you may stop it when you don't have to, and thereby cause another type of accident with other cars around you. Thus, there is a very profound question of what technology is most appropriate. The extent to which we can discuss automated driving to make it a social implementation is a major issue, and I believe there will be an opportunity to discuss this at future meetings.

I would like to close the meeting today. Finally, I would like to ask Deputy Director-General Hasui of Digital Agency to give us a summary. Mr. Hasui, please.

Deputy Director-General Hasui: I am Hasui from Digital Agency, . Thank you very much for your various comments during the very long time of more than two hours today. Among the various Issue related to automated driving, I received various opinions such as the basic concept of system security, the current system and operation of criminal liability, and the points to be considered for implementation in automated driving. Among them, I believe that you deepened your discussion based on the usage and organization of the word "criminal immunity." Thank you very much for your useful suggestions for the future. As I said in my explanation earlier, I understand that implementation in automated driving is required from the perspective of making a significant contribution to the reduction of traffic accidents. Based on that significance, I have once again realized the importance of deepening the discussion on the concept of system security and the ideal state of criminal liability, as you discussed today. As a result of your discussions, I believe that we are in the process of deepening our understanding of matters that have not been clarified previously, and we would like to continue our discussions carefully while receiving your continued support. Thank you very much. Thank you very much today.

Mr. Kozuka: Thank you very much, Mr. .

I would like to make a few administrative communications. Today, due to time constraints, if anyone has additional opinions, please contact the office by the end of this week.

My second question is. The materials for today's meeting will be published on the Digital Agency website at a later date. As in the past, we plan to publish the minutes on the Digital Agency website after confirming the content.

My third question is that the next Sub-Working Group is expected to be held in late March. I believe the Secretariat will contact you separately, including the agenda.

As mentioned above, I am very sorry that I have exceeded your time today. With this, we will close the third Sub-Working Group. Thank you very much.