Facebook Pixel

AV Perceptions Have Always Been Unmoored From Reality

Back to Resources

Written by: Ed Niedermeyer, Communications Director of PAVE

The first time I rode in one of Google’s self-driving vehicles, in late 2016, the world was quivering in anticipation of an imminent autonomous future. Leading technology pundits had begun proclaiming “the end of driving” a year before, and Elon Musk had started taking deposits for a “Full Self-Driving” feature that he said would enable Teslas to drive themselves across the country by 2018. Autonomania was in full effect.

So, when a sensor-bedecked Lexus SUV picked me up for a brief drive around Mountain View, my expectations were high. It wasn’t my first time in an autonomous vehicle, but it was my first time getting a ride from the pioneering Google Chauffeur program (now called Waymo), the 800 pound gorilla of AVs. The fact that my ride was very much a development vehicle, its screens and sensors hacked together with the air of a billionaire’s garage project, was the first hint that a polished consumer-grade robotaxi wasn’t as close as some seemed to think.

Riding in an autonomous vehicle inevitably inspires a sense of awe at the march of technological progress. Seeing the world outside rendered in minute detail on a monitor inside a vehicle that is driving itself through that world embodies the merging and interweaving of the digital and the physical in ways that VR and Second Life can only gesture at. At the same time, this vehicle’s hesitant and overly conservative driving style, even in Waymo’s sedate and familiar backyard, hinted that the technology still had much farther to go than the hype would have you believe.

To be clear, that 2016 ride always felt safe and competent but it also highlighted how heavily humans rely on our intuition and feel for traffic. This feeling comes down to little things, like subtle microadjustments to give extra room to parked cars, as if they might jump out at us at any moment. Even overly-cautious traffic spacing and slight hesitations can cause an AV to not quite fit into the flow of local traffic. Waymo had clearly done groundbreaking work to create a safe AV, but they still had a ways to go to make it feel like just another car on the road.

Two years later, as public enthusiasm toward AVs was cooling dramatically, I took another ride on the same route in Waymo’s latest build. The fatal crash of an Uber test vehicle in Tempe, Arizona and a number of delayed deployments had already started to bring dreams of imminent robotaxis crumbling down, and yet the technological progress was astounding. Waymo’s then-new Chrysler minivan felt much more like an actual consumer product, and it drove with an assertive and naturalistic confidence that could pass for a human.

This disconnect, between public perceptions of autonomous vehicles and the technical reality, has stuck with me. If people were regularly exposed to the true state of the technology, public sentiment in 2016 would be closer to the pessimism and skepticism we see today and we’d be growing more confident in it as it has improved in leaps and bounds. Instead, we continue to suffer from an extended hangover from the irrational exuberance of 2016, falling deeper into pessimism even as Waymo has begun offering rides in fully driverless vehicles with no safety driver.

Today, the public is deeply underwhelmed by the stunning progress in autonomous driving technology. A recent survey commissioned by Partners for Automated Vehicle Education shows that nearly three quarters of Americans think AVs are “not ready for prime time” and 48% say they “would never get in a taxi or ride-share vehicle that was being driven autonomously.” Yet the public’s intuition into the source of their pessimism is strong: 60% of respondents said they would have greater trust in AVs if they “understood better how the technology works,” and 58% said they would have greater trust in AVs if they “had a chance to experience an AV ride” firsthand.

Americans are right: AVs are not “ready for prime time” in the sense that so many of us thought they were in 2016. They’re also correct in believing that their current pessimism about AVs is as rooted in their lack of personal exposure to the technology as their overoptimism was in 2016. If most Americans had been able to receive a ride on public roads in a completely driverless Waymo, as I was lucky enough to just last fall, there’s no way you’d see the same level of skepticism about AVs you see in the polling data today.

As long as AVs remain abstract for most Americans, rather than a real product or service they can relate to on practical terms, they will remain a largely blank canvas onto which we project our more generalized hopes and fears about technology and the role it plays in our society. The issue, put as bluntly as possible, is that the only thing most Americans know about AVs is that they were supposed to be here by now but aren’t.

Except that AVs are here, just not as ubiquitously as many of us came to expect. AVs continue to slowly work their way into the fabric of our lives, whether in the form of Waymo’s robotaxis, Nuro’s grocery delivery pilots, or Cruise’s AVs keeping food banks stocked during the COVID-19 pandemic… not to mention the development vehicles from numerous other companies that continue to rack up thousands of safe testing miles on public roads. Progress continues to be slow and steady, but for good reason: the extreme levels of reliability required for safe autonomy demands profoundly rigorous testing, and nobody wants to risk the fragile public trust in this critical technology by moving too fast and, God forbid, breaking stuff.

As AVs slowly become a part of our lives, our hype hangover and fear of the unknown will fade. In her reflection on our poll data, PAVE Executive Director Tara Andringa describes how she has seen education and firsthand experience work powerfully to re-couple expectations to the reality of AVs. We will all go through our version of that process sooner or later, and our goal at PAVE is to help make that happen sooner rather than later.

May 19, 2020