Evidence submitted by AXA UK

Transcription

1 SEP. 08, 2016 PATHWAY TO DRIVERLESS CARS: PROPOSALS TO SUPPORT ADVANCED DRIVER ASSISTANCE SYSTEMS AND AUTOMATED VEHICLE TECHNOLOGY Evidence submitted by AXA UK AXA UK 1. AXA UK (AXA) is part of the AXA Group, a worldwide leader in financial services. AXA Group operates in 64 countries with over 165,000 employees and 103 million customers. 2. AXA has around 11 million customers in the UK and operates through specific operating companies AXA Insurance, AXA PPP healthcare and AXA Wealth. 3. AXA is a partner in three part-government funded trials of driverless and connected and autonomous vehicles VENTURER, UK Autodrive and FLOURISH. Executive Summary 4. AXA welcomes the government s ambition to support the emergence of automated vehicle technology and its continued investment in UK trials to test and demonstrate these systems. 5. AXA agrees that the current regulatory framework needs to be reviewed and ultimately amended in order to allow autonomous vehicles on UK roads. 6. Whilst these changes will inevitably take place over an extended period of time, the pace of change in the emergence of this technology means that it is very difficult to work to an exact time frame. For this reason AXA urges the government to adopt as flexible and agile approach as possible to regulation in this area. 7. To that end, certain elements of the process that will enable these vehicles to be adopted, such as relaxing sections of the Highway Code, should be actively pursued but not enacted until the relevant and proper safety case has been proven. Consultation Questions What we are proposing Question 1A: Do you agree with the proposal to review the regulatory framework to enable the use of advanced driver assistance systems and advanced vehicle technologies as they come to market in the UK? 1/9

2 The current regulatory framework is not equipped to deal with a situation in which the vehicle user is out of the control loop. In addition, there is currently no legal requirement for recorded data relevant to accidents involving automated driving to be made available for those (which could include the police as well as insurers) who would require access to it. The emergence of this technology is also likely to have a positive impact on the cost of motor insurance for consumers, given that the evidence so far suggests that automated vehicles are likely to have fewer accidents. Those who currently pay higher premiums such as young and inexperienced drivers and the elderly may see their premiums decrease as a direct result of a reformed regulatory framework. Question 1B: Do you agree that we should follow a rolling programme of regulatory reviews? This is undoubtedly an evolutionary process. Whilst the pace of change is considerable in this area, not all motor manufacturers are developing the technology and those that are doing so are at different stages of development and utilising a range of systems. As a consequence, it is impossible to say for certain when autonomous technology will be ready for UK roads and whether issues such as standardisation need to be taken into account during the process of wider adoption. An evolving framework will ensure that the necessary regulatory change will be agile enough to support the emerging technology. Question 1C: In the first wave of regulatory change, with the exception of insurance, should we only consider those advanced driver assistance systems or automated vehicle technologies that are likely to come to the UK market in the next 2-4 years? As above, with the pace of change associated with autonomous vehicle technology, consideration of systems likely to come to the UK market in the near future would allow for a smoother, evolving regulatory model to emerge. In addition, by failing to consider future advancements the government may unintendedly stifle innovation. It would not be in the interests of consumers or government for motor manufacturers to be significantly ahead of the emerging regulatory timetable. AXA also suggests that the government consider the merits of a legally recognised categorisation of increasing vehicle technology and a set of insurance, manufacture and marketing and sales regulation to support the sale of vehicles fitting into each category. This will ensure that car manufacturers market their vehicles correctly and insurance companies are able to build products which meet the legal requirements for that categorisation of vehicle. Question 1D: Are you aware of any upcoming advanced driver assistance systems or automated vehicle technologies which this document does not cover? Insurance A proportionate response Question 2A: Do you agree with the proposition to amend road vehicle compulsory insurance primary legislation in Part 6 of the Road Traffic Act 1988 to include product liability for automated vehicles? The Road Traffic Act is in place to protect road users and the public generally, and as more responsibility for accidents moves to what would generally be considered as a Product Liability area, it is important that the Act is suitably amended to include these covers as well. Key elements will be the mandatory provision of such cover (Products Liability is not currently compulsory in the UK), Limits Provided (many Products Liability covers 2/9

3 are limited to less that 5m when we have already seen substantial personal injury claims in excess of 20m), and the removal of the ability to include certain exclusions in such policies. The ready provision of appropriate compensation for any injured party must be the over-riding concern and objective, and this should also apply in respect of Professional Indemnity risks arising from software programming and design. In the short term, without legally recognised levels of automation, it would be most prudent to encourage an extension of the existing compulsory insurance obligations under a single motor policy, with potentially motor insurers dealing with claims in the first instance (under a statutory obligation) and then arranging recovery from the manufacturer / products liability insurer as appropriate. However, as the technology becomes more commonplace in cars and the legislative framework evolves to formally define differing levels of autonomy, AXA would expect product liability to be available as an add-on product with differing requirements (and probably costs) depending on that vehicle s level of technology. Clearly any responsibilities beyond those in common law imposed on motor insurers should be defined so as to be passed on to any responsible motor manufacturer or other party where their actions are held to have caused the accident. Although this will be a necessity to protect the driver/owner/registered keeper of the car in the event they are injured as a result of the car being used in autonomous mode, the lack of insurance experience will initially impact on pricing of such a product. It should also be noted that there is a potential for insurers who do not have links with manufacturers (such as holding their product liability covers) to be unfairly disadvantaged with potential customer detriment or restriction of the market. The consultation document refers to making the owner responsible for arranging this. However, such a statement raises the question of how an owner would be defined for the purposes referred to. For example, in the instance of a parent buying a car for a child, would the owner be the individual who purchased the car or the end user? The consultation document goes on to refer to a registered keeper which is, in all likelihood, a more reliable and enforceable definition (as it is supported by documentary proof as well as DVLA data). As part of that change the government must amend existing legislation regarding the sharing of data. Telematics data recorded by the car and owned by the manufacturer must be made available in a common format to insurers to ensure that a fair investigation can be conducted to determine liability and to ensure that we learn from failing technologies Question 2B: What, if any, other changes to the insurance framework should be considered to support use of AVT? As referred to previously in this submission, insurers and other parties will need to receive telematics data in a timely manner and in a common format in order to establish liability. It cannot be right that in the event of an accident, the only party with access to key information would be the one most likely to be responsible, i.e. the motor manufacturer. The consultation document does not appear to make any reference to the situation where a primary policy is voided or where there is a breach of warranty which would reduce an insurer s status to RTA. Currently there is a structure in place that covers situations where a policy may be invalid or a contract term breached and the new framework would need to include these situations. For example if the owner/registered keepers policy was voided by their insurer would any other insurer with a degree of liability for the claim still pick the claim up or could the products liability insurer be brought into a claim? The benefits and impacts of changing insurance for automated vehicles Question 2C: If you are an insurer, vehicle manufacturer or other organisation directly affected by these changes, what costs do you estimate your organisation will incur as a direct result of these changes? It is likely that claims processes will have to change due to the necessity of establishing whether or not the car was being used in autonomous mode including: obtaining the vehicle data to establish whether or not the driver was in control and; obtaining the vehicle data to establish whether all transfers between the car and driver were undertaken correctly. 3/9

4 establishing if the driver was responsive while out of the loop and able to take control and if not why not (incapacitated/ distracted). There would need to be an identifier for autonomous cars at both underwriting and claims stage. It is also likely that the questions asked at inception will be different. Current documentation will also need to change. In the longer term, connected and autonomous vehicles may be able to allow a more automated claims service. However, at this stage it is difficult to estimate the costs of such changes. Such estimates are further complicated by the expected evolving nature of both the technology and ultimate adoption on UK roads. Question 2D: Do you anticipate the cost of insurance products for vehicles with AVT to be higher than for conventional vehicles? Whilst it is difficult to answer this question with any certainty, AXA anticipates that, in the medium and longer term, the decrease in accidents on the road as the levels of technology increase will outweigh the cost of accidents resulting from the technology failing. However, in the short term, the lack of data available to insurers may mean that they take a risk-averse approach in the first instance. In addition, public acceptance and subsequent adoption (or non-adoption) of such vehicles may also have an effect. In addition, the technology involved is sophisticated equipment and the cost of repairing or replacing these parts is, as yet, unknown. As a consequence, in the short term, a large increase in repair costs may outweigh the cost savings from a reduction in accidents. Finally, as stated previously in this submission, the pace of change involved with autonomous vehicle technology means that there may be as yet unknown risks that will need to be priced in to insurance products. AXA is aware of the additional cyber-risk posed by increased levels of connectivity both from a data protection and hacking perspective; however it is too early to predict what impact such emerging risks may have on insurance premiums. Question 2E: Do you anticipate the introduction of vehicles with AVT to increase insurance premiums for conventional vehicles? There is potential for an initial impact on conventional vehicle insurance premiums given that insurers will be using burning cost models which are likely to include claims with AVs (especially given customer and other drivers) unfamiliarity. A person driving a conventional vehicle could potentially be in a less advantageous position following an accident because of the assumption that the automated car is not subject to any element of human error. However, AXA expects that as the number of vehicles with a high level of automation on the road increase, the likelihood of conventional vehicles having accidents will decrease. Currently, accidents often happen because there is a level of fault and failings by the drivers of both conventional vehicles. That said, the cost of replacing or repairing an autonomous vehicle is likely to be far higher than a conventional vehicle and this may therefore offset the expected lower level of claims frequencies for conventional vehicles. Question 2F: What do you estimate the costs will be to insurers, vehicle manufacturers, or other parties of providing product liability cover for automated vehicles, and why? As with motor insurance claims costs this is difficult to estimate due to the lack of claims experience. The testing which is currently underway as part of the government backed consortia may give us greater insight in to the likelihood and expected frequency of accidents involving autonomous vehicles, but without meaningful data and extensive testing this question cannot be answered. Question 2G: Do you anticipate that this cost will be passed on to the consumer 4/9

5 These costs are likely to be passed to the consumer in the form of claims experience and burning cost models used by insurers. If the bulk of costs fall within the products liability market this should have a benefit for motor customers however, if the motor insurance policy is the primary source of compensation then this will impact claims costs and ultimately premiums. If costs are picked up by Product Liability policies paid for by motor manufacturers, or simply picked up by those manufacturers themselves, then you would expect any additional costs to be built in to their pricing model for new vehicles. However, as stated above, it is hoped that such costs would be offset against the lower claims frequencies as a result of the increase in use of autonomous technologies. Failure to maintain automated vehicle technology, inappropriate use, and circumventing automated vehicle technology Question 2H: Do you agree that where a driver attempts to circumvent the automated vehicle technology, or fails to maintain the automated vehicle technology, the insurer should be able to exclude liability to the driver but not to any third parties who are injured as a result? The question does raise some of the definition issues that this submission has touched on previously in that it presupposes that the driver will also be the owner of the vehicle. Ownership is likely to be a key facet in the emergence of autonomous vehicles. However, AXA would see this working in a similar way to the existing MOT exclusion in motor policies. Currently, if the vehicle is not maintained and/or the vehicle does not have a valid MOT, insurers often exclude cover for the insured vehicle but not for any third party injury. AXA would expect that failing to maintain the autonomous vehicle technology would work in a similar manner and that insurers could, if appropriate, be entitled to void a policy, but deal with the injury claim as the relevant insurer under the terms of the MIB Current liability principles would treat this situation as the driver being the author of their own misfortune assuming the driver was responsible for circumventing technology or negligence in maintaining the car. However if the owner of the car did this and the driver/user as an innocent party was injured as result it would be difficult to justify not allowing an indemnity to the user/driver. The existing legislation that allows for insurers to recover from their insured any claims that have to be paid as a result of compulsory legislation should remain. Third party hacking Question 2I: Do you agree that in the event of 3rd party hacking of an automated vehicle, an insurer should not be able to exclude liability, as set out in the Consultation Document? Protection of road users and the public is paramount, as is a regime which gives confidence to the wider public, and this requires compensation to be available for injury almost regardless of the circumstances. On the face of it, AXA believes that a third party hacking of a vehicle should be treated, in the first instance, as no different to the current treatment in the event of third party theft. However, where the hacker cannot be traced or has caused mass damage consideration should be given as to how this may be covered - in this instance, AXA suggests that such an incident should be paid in a similar way to uninsured drivers (MIB levy) and recoveries chased by the MIB. It should be acknowledged that the potential for a vehicle or a number of vehicles to be hacked remotely with the intent to cause mass damage (potentially terrorist acts) carries with it the potential for billions of pounds worth of losses for any one incident. As such, AXA proposes that there is tight regulation over the risk of hacking when the technology is being developed. With the ability to recover costs from a manufacturer for failed cyber protection as yet an unknown, the insurance industry will seek comfort in a set of tight requirements and vigorous testing when developing cyber protection technology. Product liability and automated vehicles 5/9

6 Question 2J: Do you agree that the product liability and insurance requirements for automated vehicles should follow the normal rules on product liability with different rules depending on whether the injured party was an individual or a company? On the basis that, treating both individual and company property damaged by defective autonomous vehicles in the same way, is likely to have the desirable effect of raising product safety standards. However, this is caveated by the response below in relation to the state of the art defence. Do you agree that the product liability and insurance requirements for automated vehicles should be limited by the state of the art defence? AXA does not believe that it is appropriate for a state of the art defence to ever be available for accidents involving autonomous vehicles. There is a danger that such a defence would have the unintended consequence of encouraging manufacturers to bring inadequately tested cars to UK roads, which in turn may lead to accidents and serious injuries. Question 2K: Alternatively, should we extend insurance/liability rules specifically for automated vehicles? As mentioned in 2J above with regards to extending product liability rules, treating both individual and company property damaged by defective autonomous vehicles in the same way, is likely to have the desirable effect of raising product safety standards. AXA also believes that eventually extending insurance/liability rules specifically for automated vehicles will be inevitable. However, the answer provided at2a above, demonstrates that an evolving legislative and regulatory framework will present challenges in both the short and long term. Public sector vehicles Question 2L: Do you agree with the proposal that, with respect to automated vehicles, the public sector can continue to self-insure but, where they choose to self- insure, they would then be required to step into the insurer s position in respect of product liability damages? AXA agrees that autonomous vehicles should not alter the status quo with regard to public sector vehicles. An alternative option: a first party insurance model Question 2M: Do you agree that an alternative first party model option would not be proportionate while automated vehicles represent a small proportion of the fleet? Whilst the first party model would offer some benefits in terms of prompt settlement and avoidance of third party costs designed to inflate the claim overall, it is difficult to see how this would be a practicable model for insurers to follow, particularly when automated vehicles form only a very small part of the overall fleet. Question 2N: What do you anticipate the cost of implementing a first party insurance model would be? Whilst difficult to quantify, the cost is likely to be far higher than the other suggested model. Paying out claims up front, irrespective of fault and then recovering is generally a slow process and disputes over liability is likely to add unnecessary expenses to the claims handling cost which will ultimately be passed onto the customer. 6/9

7 In addition, creating a commercial relationship where the manufacturer and insurer enter into an agreement to provide products that cover both the car and the driver would have a detrimental effect on the free operation of the insurance market. Next steps Question 2O: Do you have data to support your answers on insurance for automated vehicles? Highway Code and Construction and Use Regulations Highway Code ADAS guidance Question 3A: What are your views on amending the text of the Highway Code in a way that would clarify rule: 150, related to use of driver assistance systems and distraction? 160, relating to driving with both hands on the wheel? AXA believes that both rules will inevitably need clarifying in the future. The government has stated its preference for people to use automated driving technology to free up time on their journey. As such a review of the Highway Code will be fundamental in allowing the next stage of autonomous vehicles on UK roads. Much more testing needs to be done to ascertain individuals capability of resuming control of the vehicle in certain timeframes and conditions in the event that they are relying on the vehicle s systems and traveling without hands on the wheel. Until such testing has been satisfactorily completed, no changes to the above rules should be made. Enabling platooning Question 3B: Do you agree with the proposition to allow platooning by relaxing Highway Code rule 126 (which recommends a 2 second gap between vehicles)? In September 2015, AXA published its report The Future of Driverless Haualge. 1 AXA believes that automated technology has the potential to transform the haulage industry, with very significant implications for the UK s roads, in terms of safety and congestion, for its environment, businesses and the UK economy as a whole. In commissioning this economic modelling, we wanted to discover the financial impacts of introducing driverless haulage fleets. The results confirmed our suspicion that automated freight will not only be much more efficient and make the roads safer for other users, particularly for those travelling at night; it will also reduce the prices of the end products that consumers buy. Question 3C: What, if any, other restrictions should be considered regarding use of platooning technologies, and why? AXA believes that, certainly in the short-term, platooning should be geo-fenced and, initially, only used at times of light traffic (for example at night) in order to properly assess the economic benefits, such as reduced fuel consumption, in an environment that does not overly impact on other road users. The government should also consider the requirements placed on drivers within the platoon to ensure that they are able to take back control of their individual vehicle where necessary /9

8 Freeing the driver to make use of the automated vehicle Question 3D: Do you agree with the proposition that specific and implied driver distraction restrictions are not relaxed at this time? As above, whilst in the longer term such a relaxation will be necessary to ensure take up and adoption of such vehicles in the shorter term vehicles will still need the driver to take back control if and when required. Construction and Use Regulations Remote control parking Question 3E: Do you agree with the proposed approach to enable remote control parking by clarifying: Regulation 104 (the driver should be in a position to be able to control the vehicle)? Regulation 107 (switching off the engine when the vehicle is not attended)? AXA agrees with the proposal to clarify Regulation 104 by adding a statement that a driver meets this requirement even if he is not in the driving seat, as long as he has the ability to control the vehicle through a hand-held device. AXA agrees with the proposal to clarify Regulation 107 by adding a statement that an individual, though not in the driving seat, is deemed to be attending the vehicle when they are driving it, or are about to drive it, using a handheld device. Regulation 110 (not using hand-held mobile phones while driving)? The question raises the issue of the definition of driving. AXA supports such an amendment for the purposes of enabling remote control parking only at this stage. For all other aspects of driving, AXA would advocate that the effects of using a mobile phone while driving are proven. Until the safety case is proven to allow a relaxing of elements of the Highway Code, as above, a driver should not be allowed to use a hand-held device whilst legally being deemed to be in control of the vehicle. However, for autonomous vehicles to be ultimately adopted, in the event that the vehicle is deemed to be in control, the human passenger should be free to use a telephone, handheld device, book etc. Motorway assist Question 3F: What are your views on amending Regulation 109 to allow drivers to view TV/display screens displaying information that is not related to the driving task, while driving? The question raises the issue of the definition of driving. The status quo should not allow for such a relaxation, as the driver is legally deemed to be in control of the vehicles and, therefore, potentially liable in the event of an accident occurring. However, should the government amend regulations to make the vehicle the driver it would seem reasonable to also relax Regulation 109 for the periods of time that the vehicle is deemed to be in control. Therefore, as above, for autonomous vehicles to be ultimately adopted, in the event that the vehicle is deemed to be in control, the human passenger should be free to use a telephone, handheld device, book etc. 8/9

9 The benefits and impacts of ADAS Question 3G: Do you have any data or evidence of the safety benefits of these advanced driver assistance systems? Please see answer to 3I below. Question 3H: Are there any other, non-safety, impacts (including costs) of ADAS, which we have not covered in this consultation document? The consultation document has not considered the potential cost of repairing or recalibrating the relevant equipment. In addition, it has not considered the impact on the MOT and whether, for example, software updates will be made during the MOT, or equivalent process, or by individuals/companies (dependent on ownership models). Question 3I: Please supply any data to support your answers. Insurers are only at the very early stages of quantifying as an industry although data does appear to support the findings that Thatcham Research reports 2. There is particularly a notable reduction in third party personal injury (TPPI) and total and permanent disability (TPD) claims but it should be noted that this is only on cars fitted with AEB as standard and doesn t include other types of ADAS /9