Although they sound like the stuff of the future, manufacturers have been working on developing automated cars for decades. Advocates say such vehicles could help alleviate traffic, prevent accidents and reduce emissions. But as Steven Shladover of the University of California, Berkeley’s PATH program argues in the June Scientific American, industry and media have oversold the promise of so-called “self-driving” cars. Automated cars are coming—and that’s a good thing—but they will be different from what the hype suggests.

In a colloquium presented at Halmstad University in Sweden Shladover elaborates on what automated cars can realistically do, how they can most effectively and safely be deployed, why these vehicles are worth pursuing and what the main risks are to their deployment. “Automated driving is not an end in itself,” Shladover says during the lecture, but instead “it’s a tool for solving transportation problems.”