Why the robotaxis of the future must be more than robots

Driverless cars have been the stuff of tech- and car-industry promises, as well as silver-screen appearances, for years. But when will autonomous vehicles finally merge onto highways at a steady clip? And, when they do, will they eliminate traffic and accidents and slash carbon emissions as promised?

It will require more than good hardware, experts say.

“We still cannot figure out how to make cars see the world around them to be able to make the right decisions,” says Mary “Missy” Cummings, professor of electrical and computer engineering at Duke University’s Pratt School of Engineering.

In one example, Cummings says, an autonomous “car got stuck behind an orange construction cone, and it couldn’t figure out what to do to get around it.”

So, there are issues that need fixing.

Cummings joins Michal Lev-Ram and Brian O’Keefe, the hosts of Fortune Brainstorm, to discuss the challenges of developing self-driving vehicles. Each episode of the podcast explores how technology is reshaping our world.

Also on the show is Laura Major, CTO of Motional and author of What to Expect When You’re Expecting Robots. Motional is a joint venture between Hyundai and Aptiv, which are partnering with Lyft to deploy driverless robotaxis.

“It has to provide all the functionality that a passenger expects from a taxi driver,” Major says. “That includes not only understanding its environment and navigating that, but also all of the customer assistance and the route information that we’ve come to expect as passengers in a taxi.”

Rounding out the show is Robert Seamans, associate professor at NYU’s Stern School of Business, to discuss the gray areas that are cropping up on the regulatory front when it comes to autonomous vehicles.

Subscribe to Fortune Daily to get essential business stories straight to your inbox each morning.