Trust is a big deal when it comes to autonomous cars, but the discussion to this point has been almost entirely one-sided: Can humans trust driverless vehicles? But developments at two US-based companies last week — on-demand ride service Uber's plan to put driverless Volvo sport-utility vehicles on the streets of the US city of Pittsburgh, and robotics startup NuTonomy's deployment of driverless taxicabs in the tech-centric One-North district of Singapore — prompt another question: Can driverless cars trust human passengers?
As any taxi driver will readily attest, there's more to the job than simply driving the car. Drivers may be called upon to apply fuzzy logic to "translate" slurred or syntactically incorrect destination requests, or even play "destination Pictionary" with riders for whom the language gap proves insurmountable. Even with technological advances like natural language processing and real-time translation, how can a car's passenger interface assist the passenger who, for instance, doesn't realise there are two "65th and Broadway" intersections in New York City, six miles apart? And just how will the taxi of tomorrow react to the breathless movie hero who leaps in, points, and yells, "Follow that car!"?
Human drivers also serve other, distinctly human functions — reacting appropriately when a passenger leaves a duffel bag full of cash on the back seat, or simply closing the rear door when a drunken passenger fails to oblige. And what will a driverless car do when a passenger vomits — or worse — in the back seat? The needs are apparent, but the solutions won't be simple. For now, NuTonomy is shying away from specifics. Said the company in a statement: "We are working to implement additional systems that will ensure the comfort, safety, and convenience of all passengers throughout their journey."
The company plans to introduce Singapore-wide driverless taxi service by 2018.
The presence of a taxi driver also dissuades a variety of illicit passenger behavior, including vandalism, drug use, and, of course, self-expression of a sexual nature. During NuTonomy's Singapore taxi test, says the company, an engineer will ride along "to observe system performance and assume control if needed to ensure passenger comfort and safety." Eventually, though, it will be just car and passenger. Are the cars ready for responsibility?
"None of these problems require particularly high-tech solutions", says Dr Richard Alan Peters, a professor of electrical engineering and computer science at Vanderbilt University in the US. Peters, who serves as the chief technical officer for the artificial intelligence software company Universal Robotics, suggests that driverless taxicabs could use features like automatic door closers and cabin sensors to ensure seatbelt use or tattle on smokers who light up in the car. And some tasks formerly undertaken by a human driver — discovering a forgotten parcel or a pool of vomitus, for instance — could fall to customers, "who would then alert the car", says Peters. The question, therefore, may not be whether the cars are ready for the responsibility, but whether passengers are ready for it.
With its Singapore public-road test, NuTonomy is deferring the harder questions by taking a baby-steps approach to the concept. The program is designed more as a stress-test for the autonomous-car software than for the driverless-taxi business model. For now, the company's autonomous cars offer free rides within an area of just 2.6 square miles, stopping only at set pickup and drop-off spots. And in a city-state that imposes an £1,100 fine for spitting on the sidewalk, overt passenger misbehaviour seems less than likely. But, erring on the side of caution, NuTonomy passenger participation during the trial is an invitation-only affair.
For Uber's Pittsburgh field trial, the ride-on-demand service has worked with Volvo Cars to create a small fleet of modified XC90 sport-utility vehicles. The companies have been testing the vehicles on Pittsburgh streets since May, and by the end of the year, the Chinese-owned Swedish carmaker will put 100 of the sensor-laden SUVs in Uber's hands. "Customers will request cars the normal way, via Uber's app, and will be paired with a driverless car at random", an Uber spokeswoman told the BBC. "Trips will be free for the time being, rather than the standard local rate of $1.30 [£0.98] per mile." The catch? The autonomous XC90s are speed-limited to a poky 25mph. And like the NuTonomy test cars, Uber's Volvo prototypes will have a killjoy engineer observer in the front seat.
Further down the road, the challenges of serving a diverse population — and supervising its behaviour — won't be exclusive to driverless taxicabs. Autonomous-mobility researchers, including the team behind the Google Self-Driving Car, have envisioned uses for the technology that extend well beyond short hops with clear-voiced, smartphone-wielding professionals and into trickier territory, such as transporting an elderly rider to a medical appointment, running errands with a visually impared adult, or ferrying a small child to school.
It's safe to suggest driverless-car technology won't be fully mature until it can monitor and manage the deportment of a typical 9-year-old schoolboy. And that, as any Year 4 teacher will attest, requires a lot of processing power.
If you would like to comment on this or anything else you have seen on BBC Autos, head over to our Facebook page or message us on Twitter.
And if you liked this story, sign up for the weekly bbc.com features newsletter, called “If You Only Read 6 Things This Week”. A handpicked selection of stories from BBC Autos, Future, Earth, Culture, Capital and Travel, delivered to your inbox every Friday.