When your autonomous vehicle goes autocratic.

University of Texas at Austin

The March3-9 print edition of The Economist had a great special report on autonomous vehicles. Actually “great” as a modifier to Economist special article is redundant; all their reports are worth the read and this one was no exception. Tom Standage does a masterful job of driving the reader through the future landscape of such vehicles, suggesting they’ll change the world as the automobile itself has done over the century just past. He explores the technical issues; the emerging competition among carmakers, IT leaders, startups, and ride-hailing services such as Uber and Lyft; the awkward aspects of the transition period we’re entering; and much more.

That “much more” includes an inventory of foreseen and unforeseen consequences: reductions in deaths and injuries (and reductions in organ donations); reductions in congestion and the need for urban parking acreage, associated with the rise of ride-sharing; the commercial world arriving at our doorsteps; cars that could be reshaped and outfitted as workout platforms, beauty salons, etc.; and additional infringements on privacy; possible new forms of segregation, reduced access and much more.

The report even acknowledges that meteorology, especially snowfall (blanketing lane markers and other features helpful to onboard computers busily identifying automobile position) poses challenges that have yet to be addressed.

A lot to process!

As thorough as the report is, however, it suffers from a blind spot – the same human blind spot that shapes our policies for dealing with weather hazards. That blind spot? A failure to learn from experience (following the aviation industry), and a reliance on emergency evacuation in the face of weather hazard as opposed to vigorous attention to land use and building codes that would make home the safest place to be, and shelter-in-place the preferred hazard response.

Here are three realities (or social preferences, whichever nomenclature you prefer). First, the transition to autonomous vehicles will occur on a time scale short compared with the lifetime of the nation’s building stock. That is, we will find ourselves in Autonomous Vehicle World before it will be safe to shelter in place, no matter how vigorously we might move to replace or retrofit building stocks with something safer. As a future hurricane approaches shore, thousands/millions of people in the path will still need to evacuate.

Second, current emergency response favors “mandatory” evacuation not just because of the short-term physical hazard posed to people in harm’s way, but also because an empty city is easier to manage during extreme weather than one teeming with people all facing a variety of special needs (for food, water power; emergency medical attention; protection from looters, etc.) Artificial intelligence will make it technically possible for emergency managers to do this. They might in principle take over control of vehicles during evacuations; ensure that vehicles don’t evacuate with only a single driver, but also pick up a couple of others who might need a ride, etc.

This concern might seem fanciful (“Americans would never allow such an infringement on their basic liberties”) but for the fact that autonomous vehicles are likely to be accompanied by widespread, nearly universal ridesharing, and a substantial drop in the number of vehicles on the road and a change in the ownership of those vehicles. Facing a need to evacuate, what will be the pricing structure for ridesharing vehicles? Barring regulation, what is to prevent the price-gouging already visible during peak commuter traffic, or evident in home and building repairs following a natural disaster, such as one of the hurricanes or wildfires the U.S. suffered during autumn of 2017? Will price and ability to pay alone determine access to vehicles? What are the implications for social equity in this arrangement? Would access instead be first-come, first-served? That has drawbacks too. Will ridesharing vehicles be commandeered to dump passengers off at nearby shelters (versus, say, a more distant home of a family member) to allow return visits to the hazard zone?

The Economist report suggests we’ll be better off individually and as a society to the extent we work out these and related policy issues in advance, versus belatedly discovering problems in the aftermath of a crisis.

It’s not too soon to begin thinking these problems through. Perhaps the process might spur efforts to build resilience in the built environment and critical infrastructure.