Ground Truth Autonomy: Time for a Reset on How We Talk About “Driver-Assist” Technology and “Autonomous Vehicles”
This column is republished from Ground Truth Autonomy.
Self-driving. Fully driverless. Semi-autonomous. ADS. ADAS. Automated. Fully automated self-driving. A lot of terminology is tossed around by journalists, policymakers, analysts and the auto industry itself to describe inherently different vehicle technologies and consumer experiences with driver-assist technology-equipped vehicles and autonomous vehicles (commonly referred to as “AVs”). Understandably, consumers are confused. A recent poll found that a majority of Americans have no familiarity with the often-cited SAE Levels of Driving Automation, and not enough Americans understand that even sophisticated driver-assist features do not make a vehicle “self-driving” and instead require an attentive human behind the steering wheel at all times.
It’s time for a reset in how we talk about automated driving technology.
There are two distinct kinds of automated driving capability that can be integrated into vehicles: first, technology features that assist a human driver but do not take control or responsibility for the driving; and second, systems that do take responsibility for the driving and therefore do not require a human to monitor the vehicle or do any driving at all. The former is an advanced driver assistance system (ADAS) or known simply as “driver-assist” technology. The latter describes autonomous, or “self-driving”, vehicles.
To put the term ‘self-driving’ in context, it became popular based on the pioneering work of many of the early leaders in the field of robotics and automation. The term was coined as a consumer-friendly way to talk about autonomous vehicle operation, a very complex technology, and it was only applied when proven out through the capabilities of the technology for fully autonomous operation, not just applied as a branding exercise.
Leading members of the AV industry are now uniting behind this clear delineation of “self-driving” or “autonomous” vehicles as distinct from ADAS-equipped vehicles in order to stop the conflation of ADAS with AVs. It cannot be stated enough: vehicles with driver-assist technology should not be referred to as “autonomous” or “self-driving” vehicles.
What’s an autonomous or self-driving vehicle? An AV performs the entire driving task, and the people or packages in the vehicle are just along for the ride. AVs do not require human operators, not even to serve as a backup driver, since they have built-in capability to revert to a safe state, such as pulling over, if necessary.
Before driving autonomously on public roads, companies developing AVs make sure the technology undergoes extensive development and testing in computer simulation and on test tracks. AVs are then transitioned to public roads for testing with highly-trained safety drivers remaining behind the wheel throughout the technology development process. It isn’t until the rigorous testing cycle is completed and the AVs demonstrate safe operations without any human participation that they are made available for public use.
But, what about vehicles equipped with advanced driver assistance? Vehicles with driver-assist features, like GM’s Super Cruise and Tesla’s Autopilot, are available for purchase by individuals today. They require a licensed, attentive human driver at all times and do not have the capability to take responsibility for vehicle operations in all situations. Common examples of driver-assist features include lane-keeping assistance, automatic emergency braking, adaptive cruise control, and forward-collision warning. Make no mistake, a human driver is still required for vehicle operation in ADAS-equipped vehicles.
Driver-assist technology is meant to help drivers, but a driver is still required at all times.
Both driver-assist technology and AVs share a similar goal — to increase road safety — but they require vastly different levels of sensing and computer processing capabilities and have disparate use cases, business models, consumer experiences, and target customers. For instance, unlike the common use of driver-assist features in personal vehicles, you will first see autonomous vehicles utilized in commercial fleet applications, such as the highway trucking applications Aurora is developing, the suburban ride-hail applications Waymo is offering in the Phoenix area, and the urban ride-hailing and delivery applications Argo AI is conducting in Miami and Austin.
The way we, especially those of us building, advocating for, regulating, and reporting on autonomous vehicles, talk about ADAS and AV technology really matters. The terminology matters when a journalist reports on product announcements, policy developments and breaking news. The terminology matters when a policymaker is considering how to help make roads safer and increase mobility for constituents. Likewise, the AV and automobile industries owe it to consumers to talk about the two technologies clearly and distinctly.
Argo AI’s CTO Brett Browning published a helpful explanation of the unique attributes and capabilities for the categorization of a vehicle as “self-driving,” including the ability to not just see and detect but to listen for emergency vehicle sirens, so an autonomous vehicle can pull over as any good human driver would do.
I can’t count the number of times that I’ve heard or seen someone confuse driver-assist technology with AVs. When I remind them of the clear difference, they often say something like, “Oh, you know what I mean.” But as polls and news reports have shown, many consumers don’t know the difference, and this has a real and harmful impact on their understanding of the two different technologies.
If a consumer gets in a vehicle with driver-assist features but is under the false impression that the car is “self-driving” or will “drive itself,” that can lead to misuse of the technology. When other consumers hear about an incident of driver-assist misuse but are led to believe that the car was “self-driving,” that results in misconceived notions about autonomous vehicles that put the reputation of the technology at risk.
Lumping in ADAS with autonomous technology runs the risk of jeopardizing a future with AVs. The AV industry understands that it is offering a new technology that must earn consumers’ trust, which is why AV companies prioritize safety at all points in development, testing, and eventually, deployment. AVs are designed to offer safety and mobility benefits beyond those available from driver-assist technology, for example, by reducing drunk, impaired, and distracted driving, and providing increased independence to the visually impaired and disabled. Misconceived and misplaced concerns can hinder consumer interest and support of AVs, endangering those many potential opportunities of the technology from becoming reality.
Consumers must know the difference between driver-assist and autonomous vehicles to keep our roads safe. That’s why we’re calling on everyone — industry, journalists, regulators, policymakers, and other stakeholders –to be thoughtful and precise about these clear-cut distinctions when talking about driver-assist technologies or autonomous vehicles.
Ariel Wolf is the general counsel for the Self-Driving Coalition, an organization established to educate the public, lawmakers, and regulators on the safety and societal benefits of autonomous vehicle technology.