As early as 1958, when a Science Digest article rhapsodized about the possibility of drivers settling back with a good book while their car took over in “Electronic Drive” mode, people have been dreaming of autonomous vehicles. Now that dream is finally threatening to become a reality—and policymakers aren’t sure that they’re ready for it. California, traditionally ahead of the curve on regulation, recently held a hearing in which its state senators expressed their apprehension that the technology was evolving faster than they could regulate it.
While Elon Musk’s prediction that almost all cars produced in 10 years will be autonomous may still be overly optimistic, it’s certain that the industry is rapidly evolving, lending credence to regulators’ fears of falling behind.
Enter regulatory sandboxes.
In order to balance the twin imperatives of protecting citizens and markets from immature technologies while avoiding impeding innovation which could benefit them in the long run, regulators are increasingly playing with ideas and products in so called regulatory sandboxes, or live market environments that are nevertheless constrained by the appropriate safeguards.
The technique may be a recent addition to regulators’ toolboxes, but it has already been employed extensively in other emerging sectors where creative, forward-looking regulation is needed. In the growing FinTech industry, the UK’s Financial Conduct Authority (FCA) pioneered the world’s first sandbox, but the idea was quickly adopted by regulators in Hong Kong, Thailand, Malaysia, Singapore, Australia, and Indonesia. The FCA’s first report on how their sandbox has met its objectives considers the program a success. Among other things, the sandbox cut the time and cost of bringing products to market and provided the FCA with insights for its broader regulatory work.
The Malta Gaming Authority (MGA) is currently developing a regulatory sandbox for another much-hyped emerging sector, cryptocurrencies. The MGA, already a pioneer in consumer protection in the gaming sector, is enlisting the participation of the Central Bank of Malta, auditing titan PWC, various government entities, and crypto giant Unikrn. The project aims to make tiny Malta the world’s first state to regulate Blockchain and cryptocurrencies and is currently being accompanied by public consultations running through March 9, 2018. MGA chairman Joseph Cuschieri identified the sandbox test as a critical step in regulating a rapidly evolving technology which carries both important opportunities and significant risks.
A few jurisdictions have already toyed with the idea. Catalonia provides 200 km of test track supported by a comprehensive public-private partnership linking the auto industry with various departments of the local and regional governments, telecom representatives, and academics. Singapore established a 5-year regulatory sandbox for self-driving cars in 2017, effectively turning the whole city-state into an test zone for the technology. Japan is considering instituting a similar sandbox, with panels drawn from both the public and private sectors to determine the appropriate regulations.
These sandboxes are needed now more than ever. California issued its first licenses allowing autonomous cars to drive on public roads in 2014. These were accompanied by an extensive list of limitations, including the stipulation that a backup driver must be in the car at all times. Less than four years later, its DMV is moving forward with regulations allowing tests of self-driving vehicles without a human in the car.
Statistics provided to the California DMV suggest that the cars are ready for this next step. Google’s Waymo cars, for example, only needed their engineers to seize control once every 5596 miles in 2017. And the cars learn fast. General Motors’ autonomous cars, less advanced than Google’s, had to be manually controlled once every 235 miles in 2016; by the next year they only went off autopilot once every 1254 miles on average.
Lawmakers, however, are less keen to give driverless cars the green light. Panellists at a recent California state senate hearing raised a laundry list of issues that have to be addressed, from privacy concerns to job losses resulting from increased automation. Numerous other, less obvious, policy questions remain unanswered: how will police pull over a self-driving car? Who has liability in an accident between two autonomous vehicles? Can children ride alone in a self-driving car? There is the constant threat, as with any new technology, that the innovation could be harnessed for nefarious purposes—drug smuggling, human trafficking, terrorist attacks and the like.
Developed and regulated effectively, autonomous vehicles would counterbalance these risks by providing numerous benefits, including drastically reducing road accidents which are 94% due to human error, radically cutting emissions, taking drunk drivers off the road and returning mobility to elderly and disabled people who cannot drive.
Maybe the question regulators should be asking isn’t ‘are we ready’, but ‘how can we become ready’.