Building trust in automated vehicles is a two-way street
Illustration: Rebecca Zisser/Axios
For automated vehicles to succeed, drivers will need to be able to trust them. But it'll be just as important for cars to understand who's driving, and how they're doing, so they can share control at the right times.
Why it matters: Lack of trust is one of the biggest obstacles to the adoption of automated vehicles. But trust is a relationship — a shared acknowledgement of risk. So it’s important for both cars and drivers to understand one another’s skills and limitations.
What’s happening: Veoneer, a leading supplier of AV technologies including radar and driver-monitoring systems, is researching human-machine trust and what's needed to codify it.
- And yesterday at CES in Las Vegas, a coalition of carmakers, tech companies and safety advocates announced a plan to address public fear of AVs with a new education campaign touting the benefits of automated vehicle technology.
- Through social media, technology demos and car dealer training, they aim to spread the facts about what AVs can and can't do.
Details: Researchers at Veoneer are focused on what happens during so-called "moments of truth"— those high-risk, high-emotion situations on the road where split-second decisions are critical. They've found that widespread adoption of AVs will depend on three things:
- Drivers (and passengers) must trust that automated systems will make the right decisions.
- The systems have to discern and respond to different driver skill levels and emotions.
- This human-machine interaction has to be fast and seamless, with collaboration getting better over time.
"Right now the onus is on the driver to understand the car; there's nothing the other way around."— Annika Larsson, Veoneer human factors research specialist
At CES, Veoneer demonstrated the latest version of its "learning intelligent vehicle" (LIV) in a mock smart city environment.
- The car uses a multitude of sensors and advanced AI to get to know you as you drive together, improving safety and building trust.
- It judges your facial expressions and micro-movements in your eyes to understand what you're receptive to at each moment. If you seem distracted by a crying baby, for example, you might like some help so the car's assisted driving features would kick in more frequently.
- By tracking your cognitive load, your attentiveness and your responses, the car knows whether and when to intervene.
Ultimately, automated driving is a partnership, says Larsson. Cars need to be able to collaborate with drivers, similar to the way that pilots and air traffic controllers collaborate with airplane technology.
- That's different from the traditional thinking about AVs, where it's assumed that humans will ultimately be removed entirely from the driver's seat.
- But even the CEO of industry leader Waymo has gone on record saying autonomous cars won't be able to drive in all conditions.
The bottom line: When it comes to AVs, cars need to understand driver skills as much as humans need to trust the technology.