Driving the Future of Automobiles; Artificial Intelligence

July 2017

For those growing up in the 80s, the concept of cars driven by Artificial Intelligence (A.I.) is nothing new. Two words; Knight Rider. However, three decades ago, the idea that this technology would transition from the confines of TV screens onto the roads of reality seemed at the very least ambitious, if not near-impossible.  

Now, in 2017, it appears that dreams do come true, as the automotive industry is on the cusp of entering an era of mass produced A.I. vehicles. In fact, there are already 7 million cars on the road which utilise some form of A.I., with that number set to increase to approx. 122 million by 2025. But what is A.I. and how does it, or how will it change the way that we travel from A to B?  
What is A.I. in cars?
Defined as ‘the ability of computers to imitate intelligent human intelligence’, A.I in the context of current cars usually means a combination of Advanced Driver Assistance Systems (ADAS) and In-Vehicle Infotainment (IVI).  

The latter, IVI, broadly encompasses a scope of hardware and software which deliver in-car entertainment and information — think in-built DVD players, Bluetooth, Wi-Fi, audio systems, navigation systems and access to smartphone-enabled content (social networks, weather, traffic conditions etc.). This technology is rapidly becoming more widespread, with car manufacturers installing it almost as standard in new models — BMW iDrive, Mercedes COMAND, Nissan CONNECT and Jaguar InControl, to name a few.

ADAS, on the other hand, concentrates on making driving safer via advanced technology systems. Currently, that encompasses a broad range of activities, which includes cruise control, identification hazards, automatic parking, collision avoidance and intelligence speed adaption. What makes ADAS especially interesting, is its teetering proximity to the development of driverless cars, with the line between the two technologies continuing to blur as the industry pushes for full vehicle automation.  
So, what’s next for automotive A.I.?
In terms of IVI, there are predictions that we will eventually be able to converse with our cars. While the quick-witted retorts between K.I.I.T and Michael Knight are perhaps a step too far, we can expect increased interaction between man and machine as cars respond to and even anticipate drivers’ needs.  

The aim is to positively impact the driving experience through functions which help, rather than hinder; windscreen info projections, syncing with apps to extract data, gesture and voice recognition, driver monitoring and virtual assistance. etc. It’s expected that an eventual transition of all information to one cloud will connect all A.I. capable vehicles with each other, enabling the sharing of data between millions.  

However, it’s the potential in ADAS which has been getting techies in Silicon Valley excited, as they compete to be the first to launch a fully automated driving experience. The end game is to design and create ADAS systems capable of ‘deep learning’, which would enable cars to almost mimic the human neural systems through complex A.I. algorithms. Using sensors to collect data from static objects or other vehicles, A.I. capable cars could eventually detect and recognise hazards, and predict or react to changing environments and actions, turning drivers into an extinct species as the systems take over the wheel.  
What does this mean for automotive insurance?
As A.I. advances, it’s clear that there will be a significant impact on automotive insurance, both for insurers and consumers. The predictions are overwhelmingly positive; it’s very possible that insurance premiums will decrease, as insurers utilise the swathes of big data to build better and more accurate pictures of their customers. With the capability to assess driving habits — good and bad — insurers could set prices based on how consumers actually drive, as opposed to more traditional underwriting factors like location or profession. Good drivers could be rewarded with lower premiums, with bad drivers priced out and off the roads if they’re unable to pay the price for their dangerous behaviour.  

There’s also speculation that fraudulent insurance claims could dramatically decrease. Fake insurance claims cost the industry an estimated £3million a day, and push up premiums for all motorists. But with A.I., insurers could have access to the exact conditions of an accident. Big data would provide information on the speed of the vehicles, adverse weather conditions or similarities to other claims to flag up ‘crash for cash’ scams. Ultimately, it could result in cheaper policies for law-abiding drivers. It should be noted, however, that these savings will take time to filter through into customer premiums. Cars equipped with such technology are likely to cost more to repair or replace in the event of an accident which could adversely impact the cost of insurance.   

As to who will be liable — man or manufacturer — is one of the biggest questions in the new era of A.I and automotive insurance, and it’s one without a definitive answer. If machine, not man, is controlling the vehicles, does that absolve the human driver of responsibility? Google’s self-driving car was recognised as a legal driver in the US in February 2016, but with the caveat that a fully-licensed driver would have to sit behind the wheel at all times, and would be responsible for any accident. Google, Mercedes and Volvo have voluntarily claimed liability if their autonomous vehicles cause a collision, but this is currently not a legal ruling.  

A.I. is already significantly impacting automotive insurance, and will be a crucial factor in shaping the future of the industry. While some aspects still need further clarification (liability etc.), it’s clear that we’re on the cusp of an automotive revolution, and at Maiden Insurance Partnerships we relish the opportunities it will bring. Our innovative way of working and ability to adapt mean we can future-proof our model as technology develops, offering expert advice and expertly-led automotive programme management.