Autonomous cars lend themselves to lending themselves, proponents like to say: we'll rent them instead of owning them. That way, we'll make better use of a smaller number of cars and thus pay much less on average per passenger-mile.
If so, then your next car may well be your last—that is, if you can keep it going for the 3 to 50 years it may take before cars become fully, utterly robotic. Three if you're a techno-optimist like Sergei Brin, of Google and its famous Google Car; 50 years if you're a technopessimist, like Volkswagen research director Jürgen Leohold.
Or maybe you've already bought your last car no matter which side is right. An entire ecosystem of car sharing is aborning right now, thanks to startups like Uber and Lyft, and it's based on merely human drivers. True, sharing would be even more attractive with robots, which probably explains why Google recently invested US $348 million in Uber. The startup was recently valued at around $18 billion—half again as much as Fiat.
How much money would full-bore carsharing save? Larry Burns, the former head of research for General Motors and now the director of the Program on Sustainable Mobility at Columbia University, has estimated that we'd need just 15 percent as many cars as now. That would cut costs per mile by 75 percent in Ann Arbor, Mich., according to his calculations.
Now comes a more elaborate formula from Kyle Hill, a Los Angeles-based web designer and former financial advisor. He figures that the average car in the United States costs $12,700 a year to own, all things included (even traffic tickets). Uber would cost exactly the same, he figures, if you drove precisely 15,170 kilometers a year. Drive more, and owning is the better option; drive less, and Uber is the way to go.
How about simple smartphone apps that help people join together to share a cab? According to a study in Proceedings of the National Academy of Sciences, researchers at MIT, Cornell, and Italy's Istituto di Informatica e Telematica del Consiglio Nazionale delle Ricerche, people willing to put up with an average delay of just five minutes would be able to share a cab 95 percent of the time. The scheme would cut total, cumulative trip length by at least 40 percent.
And it works even if the taxicab is dull as dishwater—so long as the phone is smart.
Philip E. Ross is a senior editor at IEEE Spectrum. His interests include transportation, energy storage, AI, and the economic aspects of technology. He has a master's degree in international affairs from Columbia University and another, in journalism, from the University of Michigan.