The self-driving truck is no longer merely a futuristic concept. Autonomous vehicles are steadily moving from experimental programs toward commercial freight operations, raising an uncomfortable question that deserves considerably more attention than whether automation can make shipping cheaper: When an 80,000-pound tractor-trailer travels down an interstate highway without a human driver, what standard of safety should society demand?
The ethical debate begins with an important fact. Human truck drivers are not infallible. They become tired, distracted, sick, angry and occasionally reckless. They can misjudge distances, fall asleep, check phones or make poor decisions after spending hours behind the wheel. Automation offers the possibility of eliminating many of those vulnerabilities.
A computer does not become sleepy after eleven hours. It does not text its spouse, drink alcohol, experience road rage or decide that reaching a destination ten minutes earlier justifies taking a dangerous chance. Autonomous systems can continuously monitor cameras, radar and other sensors while reacting faster than a human under some circumstances.
If autonomous trucks eventually prove substantially safer than human-driven trucks, there is an ethical argument for deploying them. Refusing safer technology simply because society feels more comfortable with familiar risks would itself deserve scrutiny.
But “eventually” and “prove” are doing considerable work in that sentence.
Technology companies have an obvious financial incentive to move quickly. Trucking is enormously expensive, and drivers represent both a major cost and an operational constraint. A truck that does not require sleep, meal breaks or federally mandated hours-of-service limitations could fundamentally change freight economics.
That economic opportunity cannot be allowed to determine what constitutes acceptable safety.
An autonomous passenger vehicle malfunctioning is dangerous. An autonomous tractor-trailer malfunctioning at highway speed can be catastrophic. Trucks possess enormous mass, require substantial stopping distances and operate in environments containing motorcycles, passenger vehicles, construction workers, pedestrians and emergency responders. The margin for error must therefore be extraordinarily small.
This creates a difficult ethical standard. Must autonomous trucks merely be as safe as human drivers? Or should they be demonstrably safer?
The latter standard makes more sense.
Society tolerates some degree of human error because eliminating it is impossible. That does not mean Americans should automatically accept an equivalent level of machine error when deliberately introducing a new technology. Autonomous trucking companies should have to demonstrate that their systems materially reduce serious crashes and fatalities under realistic operating conditions.
Testing also cannot be confined to ideal highways and favorable weather. America has snowstorms, black ice, construction zones, blown tires, debris, confusing lane markings and drivers who behave irrationally. Police officers manually direct traffic around accidents. Firefighters unexpectedly close lanes. Children chase objects into roads. Motorists cut directly in front of trucks without understanding stopping distance.
The real world specializes in situations engineers did not anticipate.
Then comes perhaps the most important ethical question: accountability.
When a human truck driver causes an accident through negligence, responsibility can be investigated through an established system involving the driver, trucking company, insurers and courts. But who is responsible when an autonomous truck makes the fatal decision?
Is it the trucking company? The manufacturer? The software developer? The sensor supplier? The engineer who wrote an algorithm? The company that supplied training data? Or does responsibility disappear into an impenetrable technological chain in which everyone contributed but nobody is accountable?
That cannot be permitted.
Conservatives should resist both extremes surrounding autonomous trucking. Government should not prohibit innovation simply to preserve an existing industry or employment structure. America cannot remain technologically competitive by regulating every disruptive invention out of existence.
But conservatives should be equally suspicious of corporations attempting to privatize profits while socializing technological risk. If a company removes the driver and places an algorithm behind the wheel, that company must accept meaningful responsibility for what the algorithm does.
Regulation should therefore concentrate on outcomes rather than dictating technology. Government should establish demanding, transparent safety benchmarks while allowing private companies to determine how best to meet them. Crash data involving autonomous trucks should be publicly available in meaningful form. Independent investigators should have access to the information necessary to reconstruct failures. Cybersecurity standards should be rigorous because a compromised autonomous truck could become something far more dangerous than a malfunctioning vehicle.
There should also be a clearly identifiable party legally responsible for the operation of every driverless commercial vehicle.
The trucking profession deserves consideration as well. Automation could eventually displace some drivers, particularly on predictable long-haul interstate routes. But protecting jobs cannot become an argument for rejecting technology that saves lives. The automobile displaced entire professions surrounding horses. Automation transformed factories. Computers eliminated occupations while creating others.
Economic disruption is painful, but technological progress cannot ethically be measured solely by whether existing jobs survive.
Safety, however, is different.
Human beings should never become involuntary test subjects for inadequately proven autonomous systems simply because investors, manufacturers or freight companies see enormous economic potential.
Self-driving trucks may eventually make American highways safer. They may reduce fatigue-related crashes, improve freight efficiency and solve persistent labor shortages. There is nothing inherently unethical about replacing a human driver with a machine.
The ethical failure would come from replacing human judgment before machine judgment has earned our trust.
The question should never be whether autonomous trucks are impressive enough to deploy. It should be whether they are safe enough that we would willingly put our families in the car beside one.
Until the answer is convincingly yes, nobody should be in a hurry to remove the driver.