<<

AUTONOMOUS – REWRITING THE INSURANCE SATNAV

You have reached your destination – not yet! OLDER THAN YOU THINK The fascination associated with autonomous vehicles is not new. In 1925 the Linrrican Wonder travelled the streets of New York City with crowds amazed at the appearance of a that appeared to be driving itself. In fact it was radio controlled from a separate vehicle that accompanied it on its journey.

From our contemporary position we shouldn’t lose sight of the Figure 1 provides a selective picture of key developments in fact that radio in itself was relatively new and wireless control autonomous vehicles since. Like any ‘recent’ technology, would have been a novel, if not jaw-dropping experience. Radio autonomous vehicles have had a longer history than many was the cutting-edge technology of its day having an almost would realize, have been through many ‘false starts’ or iterations magical feel. A driverless vehicle in any form was extraordinary, (depending on your view) and have an accelerating pace of however it was achieved, and the fact that this test took place development after a slow beginning. almost 100 years ago is itself surprising. The schematic in

Figure 1 – Development Milestones – Autonomous Vehicles

“Any sufficiently advanced technology is indistinguishable from magic.” Arthur C Clarke

1INSURER AUTONOMOUS CARS – REWRITING THE INSURANCE SATNAV |

1 The next significant shift took place in 1958 Nebraska and a and used neural networks in the computer logic required to joint initiative involving the State, RCA Labs and . navigate complex, unpredictable terrain. Between 1991 and This particular experiment with autonomy actually involved 1997 the US Department of Transportation aimed to develop wires and lights – infrastructure outside the vehicle embedded standardized commercially viable technologies and ran both in the surroundings used to impart control. At this stage public segregated and non-segregated (i.e. mixed autonomous/non- highways officials were helping to shape the emerging systems autonomous traffic) trials. Projects ‘went military’ in the 2000s and so the external environment became an intrinsic part of as off-road and group-based capabilities were tested in a series any solution (in-road or roadside systems were seen as key of sponsored US Army challenges. By now, technologies had to control and safety). While General Motors hypothesized an moved primarily ‘in car’. ‘electronic guide system’ in vehicle, during the 1960s, highway So, there have been many periodic trials driven primarily by infrastructure was seen to work together with any system public project work with the aim of generating significant public and the road systems essentially controlled the car. In the benefits (as highlighted above) or military enhancement; and UK embedded cables were used in a test system run by the the expansion of driverless tech in the fields of public transport Transport Research Laboratory (TRL) and resembled something (e.g. Docklands Light Railway in London or the Parkside Shuttle like the toy slot car race tracks popular at the time. Such Bus in Netherlands) and commerce (e.g. mining) has pointed to initiatives were driven by potential public service benefits with a tantalizing future vision of a more efficient, cleaner, de-risked the TRL project envisioning an increase in road capacity of 50% world. and a reduction of around 40% in accidents. The baton for research was carried by governmental and academic institutions for the remainder of the century. Preliminary research into the intelligent automated logic needed for autonomous cars was conducted at the University of Illinois in the early part of the 1970s; in Europe, EUREKA conducted the €749 million Prometheus Project on autonomous vehicles from 1987 to 1995; In the US the off-road ALV was developed

1INSURER AUTONOMOUS CARS – REWRITING THE INSURANCE SATNAV |

2 SHIFTING GEAR A significant ‘sea change’ took place around the start of the 2010s. At this point, with significant falls in hardware costs and increases in processing power, established commercial manufacturers had woken up to the possible business benefits of intelligent technologies.

Most earmarked specific budgets for ‘connected’ and of advanced driver assist systems (ADAS) – this would include autonomous technologies as the development of on-board such innovations as Adaptive , Autonomous computer management and its potential for future application Emergency Braking, Parking Assist, Lane Changing Support and entered mainstream commercial thinking. They were joined Blind Spot Monitoring. by ‘pure tech’ Google already extensively using GPS and data The benefits of Electronic Stability Control (ESC) has proved processing technologies as part of its street-view-imaging compelling. Initial data for 2007 – 2009 demonstrated a initiatives (more of this later). correlation between the adoption of ESC and the % Lives Saved Much of the car technology we now take for granted has (see Figure 2 below) and ESC is now compulsory in new US and emerged from these and related projects under the generic title EU vehicles. Consider the following figures: -

• In the United States, the Insurance Institute for Highway Safety found that ESC could help avoid Figure 2 - US Fatalities and Lives Saved 41 per cent of single-vehicle collisions while a study by the US National Highway Traffic Safety Administration shows that ESC could cut down the number of single-car collisions by 35% (source – www.tc.gc.ca). • A Department for Transport study in the UK concluded that vehicles equipped with ESC are 25% less likely to be involved in a fatal accident than those without (source – www.theaa.com). • ESC is expected to prevent c3,000 fatalities (-14%), and c50,000 injuries (-6%) per year. Speed Alert (with active gas pedal) (-5%), eCall (-4%) and Lane Keeping Support (-3%) also deliver projected reductions in fatalities. In parallel these applications also have the potential to reduce congestion, as around 15% of all congestion in Europe is due to accidents (source – www. ec.europa.eu).

Whichever figures you accept, the impact of selected intelligent reach potential). Initially driven by commercial fleet operators, technologies is significant and driving much of the components with the accelerating pace of change, has now (and thinking) that underpin autonomy. developed into the wider concept of the ‘connected vehicle’ and underpins Usage Based Insurance (UBI) initiatives such as Alongside the development of the core engineering, the pay as you drive (PAYD) pay how you drive (PHYD) programs application of data and the wider proliferation of on-board in personal auto. These programs are driven by two key computing has spurred the introduction and maturing of objectives: connectivity leading to, initially, Telematics technologies and, latterly, a significant contribution to the emergence of the • Risk assessment/management – taking on higher risks in a ‘’. controlled manner Telematics appeared on Gartner’s Hype-cycle reports towards • Claims reduction – education and feedback to improve driver the end of the 2000s as one of the few potentially ‘transformative’ perception and on-road behavior technologies of the time (but a timeline of 10 years or more to

1INSURER AUTONOMOUS CARS – REWRITING THE INSURANCE SATNAV |

3 Commercially, in personal auto, some consumers are attracted cost reductions (Source – Morgan Stanley/BCG, September to the potential of lower premiums for categorized standards of 2014). Furthermore, the reduction in losses will feed through to driver behavior. For some higher-risk groups it is the only way lower premiums, estimated to be between 10- 20%. These are they can access insurance from selected carriers. The growth of significant potential benefits and fundamentally enabled through smart-phone apps in this area (and other portable technologies) technology and the emergence of The Internet of Things (IoT). have also reduced supply-side costs providing greater access Gartner highlight that the insurance industry has been both a as the need for a fixed in-car solution is removed. Examples ‘slow and fast adopter’ of IoT, depending on what technologies include UK-based Insure the Box, now owned by and are being discussed. cited as a potential move by manufacturers to secure a greater share of the insurance market (source – www.postonline. The drivers for the above developments are now largely co.uk) and US-based Progressive who use the technology to commercial. Where these overlap with public service demonstrate potential savings as a means to generate quote considerations then a collaborative agenda emerges. However, enquiries via their Snapshot program (note - one of the pioneers commercial providers are now ‘on the offensive’ actively in the UK, Aviva, now adopt a similar approach through app lobbying for Governmental support as a global market for the gamification, Aviva Drive – source – www.telegraph.co.uk). technologies involved begins to take shape. This is a significant shift from the projects undertaken in previous decades. In the EU, the eCall emergency alert system is now a legislative requirement for manufacturers, extending the ‘reach’ of UBI into Numerous, incremental changes in on-board technology mainstream use (albeit for a specific non-commercial purpose). and machine connectivity are aggregating to the point where autonomy is seen as a realistic commercial proposition. Telematics featured as one of the top 10 technologies ‘with the greatest impact’ in 2012 with a prediction that 10% of the market would be written through PAYD by end of 2014 (Gartner, 2012). While the pace of growth has not been as fast as envisaged, Gartner still view Telematics (and its related technologies) as ‘Transformative’ with up to 20% of business ultimately conducted in this manner, but observe that it has now developed into an ‘Emerging’ technology with tangible applications. This is consistent with its original timeline to maturity of 10 years. More nascent, is the Internet of Things, where a wider web of connected machines is envisaged. Recent research concludes that ‘digital ecosystems’ could reduce combined operating ratios by as much as 21%, from a mixture of claims and non-claims

1INSURER AUTONOMOUS CARS – REWRITING THE INSURANCE SATNAV |

4 PRESENT STATE Analysts Celent, outline 5 principal stages of development on the evolution to a truly autonomous vehicle (1 - Manual, 2 - Emergency Alert and Intervention, 3 - Simple Autonomous, 4 - On Demand Autonomous and 5 - Autonomous Only). The leap from stage 2 to 3 (and in some trials, Stage 4, where the car largely drives itself) is where the current technology places this evolution.

By illustration, in October 2014, Tesla launched its Autopilot to vehicle test guidelines, potentially taking into account whether system in its Model S, with many elements of a truly autonomous a higher standard should be demanded of automated vehicles, experience (lane control, autonomous steering, braking and who would be responsible in the event of a collision and how to speed limit adjustment based on signals image recognition) ensure the safety of drivers and pedestrians. Bidders have been through the use of , 12 long-range sensors and forward encouraged to put forward proposals in areas such as safety, view cameras. This pushes the boundaries of Stage 4 above, reliability, how vehicles can communicate with each other and with largely autonomous driving on a selected area around San the environment around them and how such vehicles can help Francisco/Seattle. give an ageing population greater independence. Successful bidders will match fund projects with their own money and the In July 2015 the UK government launched a £20 million government expect the intelligent mobility market to be worth competitive fund for collaborative research and development £900 billion by 2025, so significant sums of money are at stake into driverless/autonomous vehicles, along with a code of and trails are designed to place the UK at the forefront of what practice for testing (source – www.gov.uk). The project will the legislators believe to be key technology of the future. involve an examination of the Highway Code and adjustments

The Lutz Pathfinder self-drive pod (a WHY IT MATTERS two-seater, electric-powered vehicle that is packed with 19 sensors, Autonomous vehicles will become more prevalent following cameras, radar and laser-guidance) patterns of development seen in past ‘jumps’ in innovation. will be tested in Milton Keynes and Coventry; Bristol will host the Venturer This will fundamentally change the way carriers relate to consortium, which includes the customers and other road users. Issues of liability will insurance group AXA; Greenwich will shift, providing challenges for the processes and systems run the Gateway scheme, focused on automated passenger shuttle vehicles used to manage claims. Data, video and voice will expand as well as autonomous valet parking exponentially requiring new tools and skills to manage and for adapted cars - led by the Transport make sense of the information. The protection of information Research Laboratory and also involving General Motors, as well as the AA and embedded in the process needs to be addressed – cyber- RAC motoring associations. crime has a new opportunity without adequate system controls.

As well as the legal and insurance implications of its legislation in this area (source – www.bbc.co.uk). Research introduction, the UK trials will assess the public’s reaction to the (Celent, 2015) has highlighted a number of issues slowing the technology – recent research undertaken by Virgin suggested pace of adoption, grouped broadly into two areas: - that 43% of the British public wouldn’t feel comfortable with the • Technological concerns – cost and complexity of cars and presence of driverless cars on the roads. A quarter of those repair supply-chain surveyed said that they would not get inside such a car. (source – www.virgin.com) • Attitudinal – mistrust, misuse and misunderstanding of what ‘autonomous’ means in practice And these consumer issues go some way to explaining why initiatives have been muted elsewhere. Some industry experts in the US have challenged the legal The move to fully autonomous is a technological challenge status of autonomous driving. The Tesla Model S has, in that manufacturers appear confident they can overcome with particular has raised questions around California state recently announcing that the company will accept full regulations when using the autopilot function. While the US liability whenever one of its cars is in autonomous mode. A was the first country to introduce legislation to permit testing clear statement of intent that removes one of the key barriers of automated vehicles and 4 US states have done so, 15 have to adoption. However, the legal issues around liability extend rejected bills related to automated driving. And elsewhere beyond manufacturers to the supporting supply-chain (who is in Europe, only and Sweden have reviewed their responsible when the quality of maintenance falls short and

1INSURER AUTONOMOUS CARS – REWRITING THE INSURANCE SATNAV |

5 triggers an incident that leads to injury or loss of life?) So, own initiative, as it believes that the regulations require it only it is likely that test cases in the courts will be needed to fully to report disengagements where drivers were justified in taking disentangle this issue and apportion legal responsibility. over, and not those where the car would have coped on its own. It decides whether this is the case through post incident Ford has publically stated that their cars are ready for the simulation. Consumer Watchdog, a California-based campaign challenge of autonomy. The company views customer group, said the report shows that self-driving cars still need a reluctance as the key barrier to adoption (source - gigaom. human driver behind the wheel com). Ford’s research and Innovation team have stated “There is no technology barrier from going where we are now to the This touches on age-old ethical debates over choice. Can a autonomous car. There are affordability issues, but the big car make ethical decisions in life or death situations and which barrier to overcome is customer acceptance.” Tellingly, Ford value-system do you apply? Research by the University of are now pushing previously ‘high-end’ technologies down into Alabama concludes that “the entire auto industry will have to mainstream models. The current view is that acceptance of try and come up with a solution that will be acceptable for all autonomy will come through the gradual acceptance of ADAS automakers, so that all autonomous vehicles react the same way components in the mass-market. when faced with scenarios involving unavoidable accidents” – (source – www.dmv.com). Consumer concerns are being tackled head-on by Google, with the announcement of the expansion of trial activity in London So, whichever side of the debate over safety you believe will during 2016 (source – bbc news). Both sides of the debate have determine if you feel progress is too slow or too fast. The trials been keen to examine Google’s incident data from its US trials of the 60s in Italy and Germany outlined earlier had stated rates and, predictably, view the statistics in different ways. of ‘autonomy’ of 94/95%. Trials in North America in 1980s/90s pushed this to 99%. No auto-based activity is 100% risk-free, In July 2015, Google announced that its test vehicles had been as the fatality statistics prove, and so the attitudinal debate involved in 14 ‘minor’ accidents since the project’s inception really rests on a judgement of ‘when is it safe enough’ or, more in 2009. Chris Urmson, the project leader, said that all of the accurately, ‘when will consumers believe it is safe enough?’ accidents were caused by humans driving other cars, and that 11 of the mishaps were rear-end collisions. "Our self-driving cars are being hit surprisingly often by other drivers who are distracted and not paying attention to the road. That’s a big motivator for us." Over the six years of the project’s existence the test vehicles had logged nearly 2 million miles on the road (source – www.nytimes). However, in January 2016, Google’s incident reporting (mandated by the California Department of Motor Vehicles) was published. According to this data, between September 2014 and November 2015, Google’s autonomous vehicles experienced 272 failures and would have crashed at least 13 times if their human test drivers had not intervened (source – www.guardian.com). The company is reporting only 69 incidents where the human driver took control of the car on their

1INSURER AUTONOMOUS CARS – REWRITING THE INSURANCE SATNAV |

6 THE ROAD AHEAD Technical issues of liability and manufacture will progress within the control of providers and legislators. Future progress really relies on the interplay between two behavioral dimensions: -

FigureDIMENSION 3 – Scenario Analysis – AutonomousKEY FACTORS Cars CAUSE/EFFECT Technology alignment • Commoditization or removal of conscious Customers look increasingly for “on demand” task through technology technology and technology enables them • Proliferation of standards and methods to interact flexibly with vehicle to common behavior patterns • Continued rise of ADAS technology • Connectivity standards

Consumer perception • Perception of ‘safety debate’ Consumers accept risks as acceptable and • Positivity towards technology in general pass control to machines; increasing % of autonomous vehicles make ‘non-acceptance’ • Willingness to forsake driving for pleasure the exception; Generation Y leverage new • Increasingly ‘opt in’ by default as Generation tech as status symbol (as with smartphones) Y are an increasing % of driving population

For Technology Alignment, these factors will either simplify the process of accessing functionality, enabling customers to devolve/ assume control easily as mood dictates or develop into a confusing set of options and methods of application, perhaps competing options that act as a barrier rather than an enabler. For Consumer Perception, these factors will either act as a spur to action, encouraging consumers to experiment with emerging options or they will act as an inhibitor as consumers retreat to what they already know. From the analysis above the following End-States emerge: -

Figure 4 – Autonomous Vehicles Scenario Analysis – End-state Summary END-STATE DIMENSIONAL DESCRIPTION SUMMARY Full throttle Technology Alignment of consumer desire to experiment and Provider ability to articulate a viable Simplifies / solution. Consumers largely divest themselves of control as concerns over safety and cost Consumers evaporate. Providers are able to build a constructive relationship with consumers through a Experiment transparent and beneficial delivery of services. Generation Y in particular are provided with on-demand autonomy that meet their needs for flexibility and responsiveness. Near miss Technology Consumers drawn to high-levels of experimentation due to promises of significant savings Confuses / in time, hassle and quality of service levels; However, projects are slow to deliver benefit Consumers and ultimately founder on confusion around the structure and clarity of solution delivery Experiment and safety concerns. Bad press arises forcing more and more of sector to turn back on technology. Consumers retreat to what they know and abandon experimentation leaving some Providers ‘high and dry’. Blind spot Technology Providers push a reasonably clear vision of the future but Consumers are slow to respond, Simplifies / largely due to concerns over control and data ownership/security issues. Providers are Consumers frustrated and friction builds between parties. Some low-scale projects emerge (mainly Ignore due to Generation Y with higher threshold to experimentation) but potential benefits remain largely unrealized. Parking lot Technology Complete misalignment as confusion over solution delivery leads Consumers to discount Confuses / possible benefits after some initial experimentation. Instead Consumers rely on adaptations Consumers of ADAS and Providers are left to pick up the pieces. Providers reluctantly exit to reassess Ignore future market state.

1INSURER AUTONOMOUS CARS – REWRITING THE INSURANCE SATNAV |

7 Figure 5 below represents this analysis diagrammatically: -

Figure 5 – Autonomous Cars Scenario Map

Whenever a new technology is introduced that fundamentally challenges accepted behavior then similar concerns arise (as those who remember email or, more recently, gesture-based interfaces, will remember). Manufacturers have gradually introduced more and more ‘intelligent’ components that ‘think’ for the driver (automated braking, parking assist, automatic handbrake, automatic headlights, electronic stabilization control and so on). The same road space now shares ‘old’ and ‘new’ tech cars. Small, frequent, iterative changes tend to gradually shift the balance over time and the Generation Y ultimately feel more comfortable with technology; so many commentators feel it is more a question of ‘when’ rather than ‘if’ these attitudinal factors can be overcome, but progress is unlikely to be fast-paced.

1INSURER AUTONOMOUS CARS – REWRITING THE INSURANCE SATNAV |

8 IMPACT ASSESSMENT In the medium-term the connectivity of vehicles and the technology (and business opportunities) that fall out of this are more likely to drive innovation.

Insurers are one set of stakeholders that are essentially reacting • The Data inhibitor – the relatively unstructured nature of to market development in related sectors. There are some ‘connected data’ prevents its easy categorization and limiting factors (inhibitors) that prevent a flexible, proactive subsequent use approach at present that include: - • The Privacy inhibitor – data ownership and fears over possible • The Legacy System inhibitor – prevent the integration of exploitation through cyber-crime act as a barrier to share processes and assimilation of data generated from UBI and usable information connected technology • The Product/Rating inhibitor – similarly, the ability to process, segment and apply information to build and issue new product and/or rating structures is limited by the underlying transactional system and the absence of workable underwriting models

As these inhibitors are removed or eroded the ability to re-shape the relationship insurers have with their WHY IT MATTERS customers shifts, particularly in two key areas: - Connectivity is already here and has been developing with ever • Risk profiling and Micro-insurance - more sophistication for the last 10 years. Why shouldn’t an insurance premium be calculated based on ever shorter The ‘smartphone generation’ will view this as the accepted periods of time, linked to on-going norm. Having a connected car will be seen as essential. changes in risk? As this generation explore connected services they will • Claims process enhancement and increasingly expect insurers to become proactive in how they cost reduction – Why shouldn’t an deal with service issues. insurer provide a proactive service driven from data collected and The removal of inhibitors (largely system-based) and the transmitted in near real-time by the claimant – deploying preferred ability to develop attractive connected products and services suppliers who guarantee costs and that leverage new capabilities will become key to competitive quality standards and evaluating advantage liability and circumstances with greater accuracy.

As Telematics/UBI and the Internet of Things merge and Generation Y start to demand services in different ways, ‘connectivity’ is increasingly likely to drive the agenda. How quickly this then morphs into an autonomous world is the $64,000 question.

1INSURER AUTONOMOUS CARS – REWRITING THE INSURANCE SATNAV |

9 REFERENCES Celent, April 2015, Insurers beware: speedbumps on the road to autonomous vehicles Gartner, July 2014, Hype Cycle for P&C Insurance, 2014 Gartner, December 2012, Top Industry Predicts 2013: The Nexus of Forces Will Drive Massive Transformation in Many Industries, 2012 Gartner, March 2012, Top 10 Technologies with the Greatest Impact for the Property and Casualty Insurance Industry, 2012 Estimating Lives Saved Annually by Electronic Stability Control, December 2011, Traffic Safety Facts – Research Note, US Department of Transportation, NHTSA, 2011 Insurance and Technology: Evolution and Revolution in a Digital World. Morgan Stanley and Boston Consulting Group Blue Paper, 8 September 2014. http://www.bbc.co.uk/news/uk-england-london-35511980 http://www.computerhistory.org/atchm/where-to-a-history-of-autonomous-vehicles/ http://www.dmv.com/blog/will-self-driving-cars-be-able-to-make-ethical-decisions-522213 http://ec.europa.eu/transport/themes/its/road/application_areas/vehicle_safety_systems_en.htm http://gigaom.com/2012/04/09/ford-is-ready-for-the-autonomous-car-are-drivers/ http://www.nytimes.com/2015/09/02/technology/personaltech/google-says-its-not-the-driverless-cars-fault- its-other-drivers.html?_r=0 http://www.postonline.co.uk/post/news/2390819/insure-the-box-sale-highlights-threat-of-manufacturers-to- motor-insurers https://www.tc.gc.ca/eng/motorvehiclesafety/tp-tp14651-vs200701-faq-742.htm http://www.telegraph.co.uk/motoring/news/10110097/Smartphone-apps-that-monitor-how-you-drive.html http://www.theaa.com/motoring_advice/electronic-stability-control.html http://www.theguardian.com/technology/2016/jan/12/google-self-driving-cars-mistakes-data-reports https://uk.news.yahoo.com/how-the-first--driverless-car--was-invented-in-britain-in-1960-093127757. html#TAIMeVa http://www.virgin.com/disruptors/driverless-cars-face-backlash-from-uk-public http://www.wired.com/2012/02/autonomous-vehicle-history

1INSURER AUTONOMOUS CARS – REWRITING THE INSURANCE SATNAV |

10 ABOUT 1INSURER 1insurer is a global provider of software-led solutions to insurers for more effective policy and claims management. Part of Carlyle Group, a global alternative asset manager with $188 billion of assets under management, the 1insurer Suite is operating in 8 of the top 10 global insurance markets across 4 continents. In every market, 1Insurer’s technology empowers Insurers to deliver better, faster results, making us the global partner of choice for locally delivered insurance solutions.

UNITED KINGDOM NORTH AMERICA AUSTRALIA

First Floor, Forum 4, Parkway 76 Batterson Park Road Level 5 Solent Business Park, Whiteley Farmington 1 Bowen Crescent TPO15 7AD TCT 06032 EMelbourne United Kingdom United States VIC 3004 E E 1insurer.com 1insurer.com+44 (0) 1489 850 111 1insurer.com860 674 2900 [email protected] [email protected] [email protected]