Fully controls all remaining aspects of driving. ·

Fully autonomous vehicles are vehicles, which are capable of
driving themselves by perceiving the environment and also making decisions
about safe and desirable routes (Gillian Yeomans, 2014). However, autonomy can
be divided into several levels in the process towards achieving fully
autonomous vehicles. The said degrees of autonomy according to the Society of
Automotive Engineers International’s new standard J3016 is shown below:

·        
Level 0 – No Automation

The human driver controls all aspects of
driving. This includes vehicles with warning or intervention systems.

·        
Level 1 – Driver Assistance

A driver assistance system controls either
steering or acceleration/deceleration by making use of information about the
driving environment and with the expectation that the human driver controls all
remaining aspects of driving.

·        
Level 2 – Partial Automation

One or more driver assistance systems control
both steering and acceleration/deceleration by making use of information about
the environment and with the expectation that the human driver controls all
remaining aspects of driving.

·        
Level 3 – Conditional Automation

An automated driving system controls all
aspects of driving with the expectation that the human driver responds to a
request to intervene.

·        
Level 4 – High Automation

An automated driving system controls all
aspects of driving even if the human driver does not respond to a request to
intervene.

·        
Level 5 – Full Automation

An automated driving system controls all
aspects of driving under all roadway and environmental conditions. A human
driver is no longer required.                                                                                                                                                                                         (SAE
International, 2016)

 

 

 

 

When looking at the current development of
autonomous vehicles, many automobile companies are taking immense measures to
become pioneers in the market both independently and in some cases through
partnerships. Partnerships here include tie-ups with other automobile companies
or even with information technology firms. Figure 1 compares the execution and
strategy of 18 major companies currently working towards materialising the idea
of autonomous vehicles. Points were given to automobile companies based on
specific criteria. The criteria for execution comprised of sales, marketing and
distribution, product capability, product quality and reliability, product
portfolio and staying power whereas vision, go-to-market strategy, partners,
production strategy and also technology were taken into account for strategy
(Sam Abuelsamid, David Alexander & Lisa Jerram, 2017).

Ford clearly looks to be the market leader in
terms of strategy with plans of bringing level 4 autonomous cars to the road by
2021 (Matt Burgess, 2017). With considerably the best execution and
significantly high points for strategy, General Motors has set a target of
rolling out driverless cars by 2019 in the USA. This large scale deployment has
been its goal since the estimated $ 1 billion acquisition of start-up Cruise
Automation in early 2016 (Alexandria Sage & Paul Lienert, 2017). Adding to
that, various other companies are also following suit, probably just falling
short due to a lack of resource and research.

 

 

 

Motor Insurance

Motor insurance is basically a policy aimed at
protecting individuals, their vehicles and other motorists against liability
through the payment of financial compensation in the case an accident or
collision occurs. One of the many types of motor insurance cover is the third
party only cover. Being a legal prerequisite, when a driver is the cause of a
collision, this cover provides compensation to other drivers for injuries and damages
to their vehicle. Also, this cover allows the passengers with the driver at
fault to claim an indemnity but does not cover any of the driver’s costs. This
cover is therefore present and considered compulsory to ensure victims of an
accident on the road are always protected in any given scenario. The third
party fire and theft cover is very much similar to the third party only cover
but also insures drivers when their vehicles are stolen or damaged by fire.
Lastly, as an extension to the third party fire and theft cover, the
comprehensive cover provides compensation to the driver at fault in case of
vehicle damage in a collision. Different companies also offer different
additional levels of insurance cover beyond the legal requirement. In the event
of an accident, the driver has to follow certain procedures to make a claim. If
the driver is the cause of a collision that led to damages or injuries to any
other party, the driver must give his own and also the vehicle owner’s
information to the said party if required. If these details were not given, the
driver at fault should make a report of the accident to the police within 24
hours and also his insurer even if there is no intention of making a claim. Likewise,
for accidents involving an uninsured party, the driver should also report the
incident to his insurer and the police. Furthermore, innocent victims of
uninsured and untraced motorists in situations of hit and run are ensured
certain forms of compensation by the Motor Insurers’ Bureau (MIB) (nidirect
government services).

Having said that, my dissertation aims at
studying the impact of the gradually improving levels of vehicle autonomy on
the motor insurance industry from varying perspectives.

 

 

 

 

 

 

 

 

Main Content (Try introduce technical terms, find data or model if
possible)

Chapter 1: Underwriting

Currently, over 90% of all road accidents are caused by driver
error (AXA, 2017). Hence, it is obvious why the majority of traditional
underwriting criteria revolves around the driver namely the number and kind of
previous accidents and also miles driven (Insurance Information Institute,
2016). With autonomous vehicles, many of the said traditional criteria will
still remain, however, greater significance will be placed on model of the vehicle
and previous accidents of specific vehicle models, as driver error will
eventually be eliminated altogether. Moreover, insurance companies will also
have to constantly be aware of progressing hardware and software improvements.
This is because understanding the level of integration of these improvements
will further aid insurers assess the risks involved. Similarly, the location
where the autonomous vehicle is driven majority of the time will also become a
far more crucial underwriting criterion. Initially, this will be to measure
rating factors like theft and amount of traffic, which may lead to higher
probability of accidents. Nonetheless, for driverless vehicle underwriting,
location needs to be considered because different areas or states place
different levels of importance on providing infrastructure suited to these
vehicles. Hence, insufficient facilities and certain natural conditions may act
as impediments to automated driving. From a varying perspective, a more
complicated underwriting criterion would be to recognise the ongoing
development of less than fully autonomous vehicles. The reason is that lack of
understanding on partial automation may lead to drivers assuming the presence
of certain functions that the vehicle may not have been designed to do in the first
place. This false sense of security can in turn result in drivers neglecting
certain driving skills causing an increase in accidents. All in all, measuring
and recording the criteria discussed above will not be a problem, as autonomous
vehicles will most probably include telematics devices. This is further
supported by the fact that The National Association of Insurance Commissioners
forecasts the use of telematics to rise up to 20% within the next 5 years in
the US. Nevertheless, privacy will surely be an issue to certain groups as data
collected by insurers may also be exposed to the possibility of hacks.
Therefore, insurance companies will need to find alternative methods to
identify, collect and analyse data and adjust policy coverage and premiums in
line with the ever changing risks involved (Namic, 2017).

 

 

 

 

Chapter 2: Regulation

This leads to the all-important question of who will bear the
liability of an accident involving autonomous vehicles. Based on the Vehicle
Technology and Aviation Bill in the UK, insurers will be considered default
liable for death, personal injury or damages from accidents caused by
autonomous vehicles in self-driving mode. For this, besides being insured, the
vehicle has to be on the government’s list of all automated vehicles in the UK.
Insurers would then be able to recover the cost of damages pay outs from
respective automobile manufacturers. In addition, the bill also states that
insurers would not be liable for certain cases including accidents or damages
caused by driver negligence, alterations to the vehicle’s operating system and
failure to install software updates (Out-Law.Com, 2017). These provisions
correspond to the government’s requirement for drivers to take out dual
insurance policies for autonomous vehicles. Under the two-in-one insurance
product, drivers will be covered for both when they are in control of the
vehicle and also when the vehicle is in the driverless mode. This way the
problem of confusion of who the not-at-fault party needs to make a claim against
will be solved allowing drivers to be properly compensated in any event of a
collision (Katie Morley, 2017).

US Senators also have recently reached a deal on a self-driving
car legislation, however, under the legislation, states were still able to set
their own rules on liability and insurance (David Shepardson, 2017). Thus, legal scholars propose that
negligence should still remain as the underlying factor when considering
liability be it driver negligence or the decision making of autonomous vehicles
given the presence of a defect in the vehicle. On the other hand, a report
issued by the American Association of Justice (AAJ) suggests that automobile
manufacturers should be fully liable for any collisions involving their
driverless vehicles (Victor Schwartz, 2017). This is also further supported by
the fact that car manufacturers namely Volvo have also agreed to accept full
liability in the case that one of its cars is in autonomous mode when an
accident occurs (Adrian Flux, 2015)

Nevertheless, to strike a balance between providing compensation
to drivers in accidents involving driverless vehicles and also not hindering
the development of autonomous vehicles in the future, legal scholars have come
up with 2 alternative liability theories. Firstly, they suggested that drivers
carry a no-fault liability insurance. In this case, while holding their own
insurance, drivers will be compensated up to a certain level regardless of who
was legally at fault in a collision. This approach was backed by the RAND Corporation
which saw it as a much more beneficial alternative to drivers as compared to
the direct shift in liability from driver to manufacturer. The next alternative
was to establish a victim compensation fund for drivers who need to make a
claim from accidents involving driverless vehicles. This way victims are able
to cover losses due to injuries and damages while at same time ensuring the
development of autonomous technology which has the potential to significantly
improve road safety is not impeded (Victor Schwartz, 2017).

Therefore, it can be concluded that liability will gradually shift
from individual drivers to manufacturers inclusive of automobile companies,
original equipment manufacturers (OEMs) and autonomous software designers
(Adrian Flux, 2015). Accidents involving autonomous vehicles but caused by
manually driven vehicles on the other hand should still comply with current
insurance liability procedures.

 

Chapter 3: Difficulty in determining the cause of accidents

Adding to that, the transition between autonomous and human-driven
vehicles in the short run will also bring about even more complex challenges to
insurers. With current driverless cars having a dual mode, it allows the human
occupant to take control of the vehicle. Such a situation calls for the
combination of both the driver’s responsibility for damages and criminal
offences and also the manufacturer’s liability for defective products
approaches (Zsolt Szalay, Tamás Tettamanti, Domokos Esztergár-Kiss, István
Varga & Cesare Bartolini, 2017). This will make it difficult to determine
if driver negligence or vehicle defect was the actual cause in the case of a
collision occurring.

Furthermore, driving also requires complex social interactions
which autonomous vehicles may not be able to fully adapt to. Autonomous
vehicles are currently being developed and tested to ensure that they are able
to react to the behaviours of pedestrians and other drivers on the road. Google
for instance has improved its cars’ software enabling driverless cars to
recognise cyclists and interpreting their hand signals. However, according to
Edwin Olson, an Associate Professor of Computer Science and Engineering at the
University of Michigan who studies Autonomy, this is just one of the thousands
of risks that may arise when driving. He explains that the presence of subtle
changes like a traffic officer in charge instead of traffic lights or even
navigating four way intersections may confuse driverless vehicles. Also, a
driver may be able to predict based on behaviour if a pedestrian using a phone
is going to stand still or abruptly cross without checking traffic but an
autonomous vehicle may not. Hence, to be ready to drive in real life scenarios,
driverless vehicles must be able to understand the environment and human
behaviour, besides being able to respond and communicate with other road users,
which may take significant effort and time to develop (Brad Plumer, 2016).

This again raises the issue for insurers to determine the party at
fault in the event of an accident. Will the vehicles incapability to adapt to
real world situations be considered a defect in the vehicle or a feature that
the vehicle did not possess in the first place? In May 2016, an accident
involving a Tesla Model S crashing into a truck that took the driver’s life
rose questions about the car’s crash-avoidance Autopilot system. According to
an industry executive, the cause of the collision may in part be the design of
the crash-avoidance system, which was designed to engage only when radar and
computer vision systems agreed that an obstacle was present. Federal
auto-safety regulators later claimed that Tesla’s Autopilot system that was
used had no defects (Guilbert Gates, Kevin Granville, John Markoff, Karl
Russell & Anjali Singhvi, 2016). Hence, in similar cases, detailed
investigation and analysis will have to be carried out by insurers to
definitively determine the actual cause of the accident. This may incur
additional costs and impact the overall cash flow of insurance companies that
generally aim to maximise their profits.

BACK TO TOP