Bad robot: Will you trust your autonomous vehicle?

Leland Teschler | Executive Editor

If you believe the hype coming out of Silicon Valley start-ups, we are on the verge of seeing highways filled with fully robotic drivers. But numerous well-chronicled accidents among semi-autonomous vehicles have brought many questions about whether people should trust artificially intelligent cars to make good decisions.

We don’t have to speculate about reactions to new driver assistance technology. The market research firm J.D. Power recently surveyed consumers on their experiences with advanced features such as blind spot warning systems and lane-centering systems. It is illuminating to read the feedback J.D. Power received. The lesson seems to be that today’s driver assistance tech is a reality check for what the future holds.

Comments coming back to J.D. Power about lane-centering were particularly noteworthy in that this feature takes control of the car’s steering – thus lane-centering is a harbinger of what may be in store when fully automatic steering debuts. “Nice try. It works sometimes,” was one comment. “Doesn’t work fast enough and makes wrong control inputs at times. Works best on highways and otherwise I’m very attentive. The system hasn’t gained my confidence.”

“I thought that the car would correct itself but it does not seem like it does,” went another. “Lane keeping is not consistent. Sometimes (rarely) the corrections are applied to the steering, but more often just get an alert in the instrument display. Unreliable. Not useful,” the driver continued.

With these sorts of comments, it is unsurprising that drivers often choose not to use lane centering. Only a bit over half (57%) of drivers reported keeping the lane-keeping function switched on all the time. About 20% say they tried it and stopped using it, never tried it, or used less than half the time.

The statistics are better for blind spot warnings, perhaps because these are alerts for the human driver rather than a system that takes control of the steering wheel. More than 80% of drivers have the warnings switched on constantly. Only about 5% don’t bother with it much of the time.

Drivers also gave J.D. Power opinions about where these advanced features needed improvement. Almost half said lane-keeping systems just weren’t accurate enough and needed better customization.

You’d have to say that so far, opinions about driver assistance functions aren’t exactly a ringing endorsement for the more complicated systems now on the drawing boards. That skepticism is reflected in the responses J.D. Powers got when it asked drivers whether they would trust fully automated self-driving cars. Even premium car owners who own vehicles likely to sport driver assistance features are skeptical of self-driving cars. More than 50% of them said they definitely or probably wouldn’t trust self-driving vehicles. An even higher percentage of people who drive ordinary cars say they wouldn’t trust robotic drivers.

All in all, indications are the automated driving features that have debuted so far aren’t exactly a big hit with the driving public. They often seem to be pretty fragile. It will take a lot of positive experiences rather than just hype to convince the public that robotic driving is an option worth paying for.