Disclaimer: The following post contains only my personal opinions and not those of my employer.
Cars, and the industry that builds, sells and services them, have changed before our eyes. We can see every day how Autonomous, Connected, Electric and Shared technologies (ACES) are remaking our cities.
Our ReachNow members already experience three of those four ideas. Every car in our fleet is a Shared asset. That removes the need for everyone in the city to have their own private car on the road or taking up a parking space. Cars in our services, along with most new cars sold today, are already Connected. Communication to the Internet cloud services allows users to find, unlock and rent our cars at any time. Many of the most exciting new cars are Electric. Before EVs can become ubiquitous most cities will need to expand public charging infrastructure.
So what about Autonomous? How soon will you be riding in style while a computer does all the driving? Right now autonomous cars are not ready for general availability. The technology is not ready. Yet.
Many companies, from automotive leaders and technology stalwarts to brand new startups, are developing autonomous car technology, yet no system is ready for all driving situations. When you buy or rent a car today, you know that car safe to drive day or night, rain or shine, so long as the driver is attentive and uses their best judgement. You can’t say that yet about an autonomous car whose driver is not a person but a program.
Some observers remain deeply skeptical about whether an AI can drive a car in all situations. That’s understandable at this point in autonomous vehicle development. Everyone wants to feel safe. These cars, when faced with the same situations that good human drivers often misjudge, must react to protect those inside and outside the car.
Even if the tech were ready, as a society we have a lot of decisions to make. What are the licensing and behavior rules for passengers in an autonomous car? How will governments handle regulating these new vehicles whether privately owned or part of a shared fleet?
I’m less concerned whether we can develop the technologies to deploy self-driving cars. GPS data and sensors in today’s cars can already help us pinpoint potholes and slick road conditions. Advancements in miniaturization of senors that enable the car to “see” are accelerating. Real time processing of the input, with ever smaller computing systems almost here. Data collected from test fleets and pilot programs will help to teach these systems how best to analyze the road around them. Soon not crashing will be the “easy” part.
I’ve spent my career building software services for consumers. In my experience, the hard hard problem isn’t what happens outside the car, but inside.
Think about your last trip in a taxi or ride share vehicle. Did you interact with the driver as you reached your destination? Was there a moment when you leaned towards the front seat, stopping your cell phone conversation or game of Candy Crush just long enough to point at the street ahead to say “Pull up right here.”
Seems simple, right? But where exactly is “right here”? Is the car going to use a vision system to interpret where you’re pointing? There’s no wheel to turn. The doors are locked for your safety, so you can’t hop out. You might be stuck repeating the same voice command over and over until the car complies.
Humans can (generally) understand what you want and need. Language, empathy and experience make it easier to interact with a person in the front seat. That’s despite the examples of bad trips we’ve all had with frustrating drivers.
Now imagine you have just broken your foot and are using crutches while it heals. A human driver would understand it will take you more time and more room to get out of the car. The best drivers will at least make the offer to come assist you. How will the autonomous car “see” your current situation? How will it handle someone who is older or has mobility issues?
Think back to the simple request, “Just pull up right here.” How would you rate the driver that ignored your request? Unless the car AI react to basic consumer commands, who will trust the driving safety of the system?
I’m not here to say that we shouldn’t build autonomous cars. Far from it. The improvements in traffic efficiency and economy of autonomous electric fleets are critical for our increasingly dense urban neighborhoods. We need autonomous cars to be part of the transportation solution and we need them soon. Still the most difficult problem we have to solve is passenger experience feels responsive, safe and secure.
Too often people think about a future self-driving car as another place to watch Netflix. Of course we’ll be able to make new use of time we’d have otherwise spent focused on the road. The passenger experience inside an autonomous car can’t only be about finding new screen time. It has to be about creating something that is responsive, equitable and empathetic no matter who might taking the next trip.