Automobiles, as an advantageous alternative to horse-drawn carriages, were designed to make people more independent and provide them with a more comfortable means of travel. But progress has often met with obstacles that arise from prejudice.

In the mid-19th-century, one of the UK’s several Locomotive Acts greatly hindered early development efforts... By demanding that cars be preceded by a man — on foot — waving a red flag. Decades later, people still thought automobiles were no match for a horse and buggy.

Meanwhile, a lot of asphalt has run under the world’s tyres. Technological evolution has brought automated systems, like sensors, parking cameras, warning calls that blare out when you veer off your lane, warnings on braking distance and emergency braking methods (ADAS systems). But these are all about supporting drivers; autonomous vehicles (AV) intend to replace them.

Currently, while major tech firms develop their own AVs and lobby to get them to circulate on public roads, most people look upon the AV with a blend of curiosity, wariness and denial. Today, the machine would replace its creator and remove him or her from their seat behind the wheel. Can that really happen? Come back in 20 years for an update.

Public opinion and the thousand-headed hydra of social networks will certainly play a role when it comes to adopting this technology — because it will exert pressure on lawmakers, but also because the public has the power to tank projects when it simply refuses to buy.

Safer, more efficient, and... uglier?
 

The benefits of AVs are clear, though even top-of-the-line models fail to impress, aesthetically speaking.

1. Safety and lower accident rates

Typically, a major benefit when AVs are introduced. Driving accidents represent a major cause of death (1.35 million a year) and a general public health issue. Drivers are almost exclusively responsible for these accidents, so removing the human factor appears to be an incontrovertible panacea. A number of studies points to a 90% increase in road safety. In many cases, setting aside negligence for now, an AV is better at detecting dangers than human beings, merely because it has more refined “senses”. Consider how one of these vehicles can spot obstacles in low- or zero-visibility circumstances by using its LIDAR or other sensors. Or how quickly it responds — AVs practically remove response times. However, we will never do away with collisions, not entirely:

a)  At a predictably long intermediate stage we’ll have AVs and human-operated vehicles sharing the road, and machines will have trouble dealing with drivers’ reactions.  Later on, there will still be outlier and unpredictable events that may cause crashes: Storms, animals racing across the road, pedestrians suddenly emerging from behind large vehicles that had obscured them from view.

b) Nor can we rule out mechanical, electric or computer failures when it comes to AVs. The current estimate is that AV programming will require something like 300 million lines of code. Even at a critical error rate of 1% we’re dealing with a considerable number of errors that may affect vehicle safety. Although complexity will grow with AVs, you already have complexity at this very moment. Several vehicles roll off the factory floor with 100 networked devices, some of them hooked up to security systems.

      2.  Accessibility - autonomous travel for disabled people (impaired mobility or vision), or people without a drivers’ license.

      3. Convenience and comfort - driving time freed up for other activities, either leisurely or work-related.

      4. Lessened environmental cost - more efficient use through ride-sharing, better speed controls and rational urban traffic will lead to lower energy production and vehicle production costs.

      5. Improved urban areas - studies point out parking space in cities will decrease by 68%, freeing up space for people.
 

Challenges to widespread use of AVs
 

Currently, most vehicles available for private use rank at level 2 (such as Teslas) or level 3 (a few Mercedes models, for example), according to the National Highway and Transportation Safety Administration (NHTSA).  Level 2 means drivers need to pay attention at all times and be ready to take over driving duties. Most reported collisions, which always get generous media coverage, arise from drivers’ mistakes — they didn’t take over when they should have.

Over the past two decades, self-driving vehicles have seen accelerated evolution thanks to investment from major tech firms such as Google, Uber, and Apple. New companies, such as Zoox, have quickly gone from prototypes to publicly available mass transportation (taxis and buses). They are currently piloting level 4 and 5 vehicles, especially on the West Coast of the US, but also in Europe (VW Id.Buzz, for one).

The future depends on the outcome of these experiments and what kind of public acceptance they will receive, but a few key challenges remain. They’re not impossible to overcome, but they will determine whether AVs become the rule, rather than exception in transportation.

       1.  Infrastructure
AVs require a 5G communications network based on clear standards, with their own allotted frequencies that will allow vehicles, traffic lanes, traffic control systems and public bodies to interact.

      2. Challenges in regulation
Legislation must evolve not only across the EU but also coordinate with other producing countries — a thorny issue that stimulates controversy.

a) AVs must comply with EU safety standards as dictated by the General Safety Regulation (EU) 2019/2144 and further include cyber security rules.
b) Testing- and development-related legislation for AVs is necessary, and some countries have begun to introduce such legislation.
c) Additionally, self-driving vehicles will have to follow the rules of the road as enforced within each national jurisdiction — say, how to approach and use roundabouts and intersections. Clearly there is an opportunity for standardization.
d) A clear definition of product responsibility. AVs being the complex product they are, involving a host of manufacturers, we need to determine who is responsible and when.  Primary liability will always be ascribed to the automotive brand.
e) Data protection and privacy: Autonomous Vehicles must abide by the GDPR, even if they are manufactured in more “permissive” jurisdictions.

3. Challenges for the insurance industry

Autonomy presupposes responsibility. We all know that.  Civil liability for autonomous vehicles is a complex issue that tests current institutions. It is a rather complex matter that the EU must legislate as thoroughly as possible.

A number of options have come under study but as yet no clear answers exist. What we do know is that traditional civil liability, which falls to the driver in the first place, is about to lose ground or, at least, needs progressive mitigation from other forms of liability. Be it an increased responsibility for vehicle risk, or bolstered product liability — both of which present limitations and intricacies of their own — a clear path is necessary so that legal security, underpinned by a mandatory insurance scheme, can go on guaranteeing much-needed social tranquillity.

New risks will demand consideration, such as cyber security or manufacturers’ liability, which will need to factor into car insurance.  Operational standards will have to change in order to guarantee access to vehicle data so that liability can be ascertained. Without such facilities, we will face a cluster of practical difficulties when it comes time to settle a claim. That could send us back to the way things were before mandatory insurance existed in Europe.

The Long and Winding Road

A lot of work lies ahead as we must inform and educate users rationally and effectively on the risk and forward motion of these new vehicles.  It won’t be easy to persuade this century’s “coachmen” that they are safe with their self-driving cars. There must be education, possibly even a dedicated AV licence. Users must be ready to take over driving duties under certain circumstances. But will they be able to? And at what cost? Think of unmonitored children traveling, or people with physical impairments; or occupants who have consumed substances that would bar them from driving.

The risks associated with cybersecurity could prove a stumbling block for public acceptance, and we’ll have to demystify the dangers of computer attacks without underhanded dismissals. Another matter attracts controversy, and it won’t be easy to settle: ethical decision parameters. Vehicle decisions will be based on predefined cases and artificial intelligence will apply these to real-world situations. Will a vehicle protect the child traveling within, or the one on the crosswalk? The elderly citizen aboard or the dog that runs in front of the car? Will it hit the imam on foot, or risk hurting the nun it is driving to the convent? And will it strike a cow in India or swerve to crash into pedestrians on the sidewalk?

Success lies ahead, but not without hurdles. It will take sensitivity and common sense — you don’t need a man marching ahead of your self-driving vehicle waving a red flag at passersby.