Monday, May 15, 2017

What autonomous cars tell us about the future of innovation

May you live in interesting times.  This is supposedly an ancient Chinese saying, replete with both opportunity and a veiled threat.  Interesting times can introduce great new innovations but can also inaugurate great conflicts.  Nowhere is this more evident than the evolving phenomena of autonomous vehicles.

Innovators across the globe, from the large (Amazon, Google) to the small are all working on autonomous vehicles.  Futurists are already predicting the demise of the long distance truck driver, whose job will be eliminated momentarily by autonomous trucks spanning long distances.  Soon, we are promised, we'll commute to work reading our morning papers - except there won't be papers - while the autonomous vehicle takes us efficiently and safely on our way.

If you are reading a hint of sarcasm in the above scenario, you are right.  We are reliving all of the hype about electric vehicles, circa 1991, when California demanded that 5% of all cars would be zero emission vehicles within 5 years.  They may reach that threshold in 2020. The promise and excitement of technology often ignores significant barriers to adoption.

Technology is the answer not the problem.

Thanks to Moore's Law and other advances in technology, we know that the bulky, unwieldy stuff welded onto prototype autonomous vehicles will be miniaturized and made more robust.  This will take time but it will happen if the markets are there.  The costs which are a bit prohibitive today will also fall.  The combinations of big data, real time analytics, sensors, on board processing, all linked to a very responsive processor connected to the engine and drive train will provide a safe and simple journey.  The technology won't be the problem.  But just like the electric vehicle, we'll discover the real barriers to innovation:  invested infrastructure, risk and regulation.

Invested Infrastructure

One of the items that slowed the adoption of the electric vehicle was the ubiquitous gas station.  When you have to carefully plan your trips to swing by electric charging stations, the overhead associated with driving an electric vehicle can be a bit overwhelming.  As battery lives improve and more charging stations are developed, driving an electric vehicle will be more predictable.  But we still face the challenge of deeply invested infrastructure that does not fully support electric vehicles.  And this is a dilemma that innovators always face - they must work within the existing infrastructure and compatibility or disrupt it.  And if they disrupt it they better have a fully operational highly capable alternative ready to go for the masses.  That was easier for Jobs and iTunes than it will be for fueling cars.

The invested infrastructure favors human drivers who are somewhat unpredictable, who must stop or yield at intersections.  The infrastructure favors drivers because we have significant investments in parking lots - where stupid cars wait on their drivers.  Autonomous cars don't require close in parking lots, because they can drive themselves or do other chores rather than sit in the lot.  Existing investments could become catastrophically obsolete when cars are smarter - this means some interests gain, while others stand to lose dramatically.


The organizations most likely to block rapid adoption of autonomous vehicles are the insurance companies.  We know what human drivers are likely to do in a car based on decades of historical data.  We know the frequency of accidents, the cost of those accidents and so forth.  With an autonomous vehicle there are several challenges.  First, where's the past data to build models on?  We don't have a lot of historical data so estimating and pricing will be difficult.  Second, who is responsible for an accident?  The vehicle, or a subsystem, like the navigation or data processor or the sensors?  Is there joint and several liability?  Third, what are the rules? Will the population revolt if autonomous cars (which are in fact robots) accidentally take a human life, ignoring Arthur Clarke's three Robotic laws? Risk is a terrible thing for insurance companies, and until they can quantify and validate the risk associated with autonomous vehicles they won't create insurance at comparable prices to cars with drivers.  No insurance, no autonomous vehicles.


So, here's a question.  Suppose you are driving in Silicon Valley in your autonomous vehicle and you are crossing political boundaries - say you are driving from Palo Alto, where autonomous vehicles have been approved, and you cross over into San Francisco or some other municipality where autonomous vehicles aren't approved.  Given the Byzantine number of local, regional, state and federal regulations and approval bodies, you may never be able to leave your local municipality in an autonomous vehicle.  Existing regulation is often a barrier to any radically new innovation, and it will be with the autonomous vehicle.  Until all the municipalities agree or California and other states provide uniform regulations for autonomous vehicles, they are interesting museum pieces.
Regulations will have to change, and when regulations change there are winners and losers.  The organizations and companies that think they have something to lose - bus drivers, truck drivers, delivery people, etc - will put pressure on their representatives to slow or stall new regulation.

Finally, what does the innovation force to occur

In reality, there's one other barrier to innovation that the autonomous vehicle signals - the required changes once the innovation is accepted.  In reality, once a few autonomous vehicles are on the road, we'd all be a lot safer, more than likely, if all the cars were autonomous.  This would lower unpredictable driving and reduce accidents.  But that means than one can't be a little pregnant.  Either there are no autonomous vehicles or they all are.  Innovations, once accepted, often cause secondary and tertiary effects, and this is another reason there is resistance to innovation.  People can't always voice their concerns but innately they know that new technologies have knock on effects, and therefore they resist new innovations because of the unknown and unexpected consequences.

The knock on effects are both knowable and unknowable with autonomous cars.  Some scientists have speculated that autonomous cars could reduce the need for stop lights, because the cars and algorithms could sequence cars more effectively.  But this sequencing and hive mentality also suggests that a simple glitch or a hack could disable thousands of vehicles simultaneously.  Imagine a huge fleet of autonomous vehicles hit with a "WannaCry" like virus that shuts them all down with no warning.  Innovation offers great benefits but also secondary and tertiary knock on effects that must be understood, especially when human life is at stake.

What this means, generally

The autonomous car has a lot of promise, but some peril as well.  As an innovation, it is a good example of the factors that will block or become barriers for innovation.  Risk, especially when someone else bears the cost (like insurers) will always be a barrier to adopting new innovations.  Regulations will be as well.  The more disruptive the innovation, the more it will upend existing infrastructure and that will cause significant grief to a large portion of the population.  As innovators we must recognize that the value and benefit of an innovation must be measured in proportion to its impact on existing infrastructure, risk and regulation, because these factors can delay adoption.  We must also understand the secondary and tertiary effects of innovation, to understand potential barriers or resistance from those who will "lose" if the new innovation is adopted.

Some innovations, Facebook as an example, didn't encounter many of these issues.  There were few existing examples except printed facebooks at colleges.  There are few regulatory issues and little knock on effects.  However, when you innovate in a space where there is a lot of history - over 100 years of  automotive usage - then you're going to encounter resistance from those who have a stake in the status quo, and from regulation that exists to manage and sustain the existing way of life.  This should lead to a conclusion:  when you choose your targets and opportunities to innovate, choose carefully, because some innovation opportunities are much easier to realize than others.
AddThis Social Bookmark Button
posted by Jeffrey Phillips at 7:30 AM


Post a Comment

<< Home