Road trips. Drive-throughs. Shopping malls. Freeways. Car chases. Road rage.

Cars changed the world in all sorts of unforeseen ways. They granted enormous personal freedom, but in return they imposed heavy costs.

People working on autonomous vehicles (AV) generally see their main benefits as mitigating those costs, notably road crashes, pollution and congestion. GM's boss, Mary Barra, likes to talk of "zero crashes, zero emissions and zero congestion." AVs, their champions argue, can offer all the advantages of cars without the drawbacks.

In particular, AVs could greatly reduce deaths and injuries from crashes. Globally, around 1.25 million people die in such crashes each year, according to the WHO; it is the leading cause of death among those ages 15 to 29. Another 20 million to 50 million people are injured. Most crashes occur in developing countries, where the arrival of autonomous vehicles is still some way off.

But evidence that AVs are safer is already building up. Waymo's vehicles have driven 4 million miles on public roads; the only crashes that they have been involved in while driving autonomously were caused by humans in other vehicles. AVs have superhuman perception and can slam on the brakes in less than a millisecond, compared with a second or so for human drivers.

But "better than human" is a low bar. People seem prepared to tolerate deaths caused by human drivers, but AVs will have to be more or less infallible. A realistic goal is a thousandfold improvement over human drivers, said Amnon Shashua of Mobileye, a maker of AV technology. That would reduce the number of road deaths in the United States each year from 40,000 to 40, a level last seen in 1900.

Building efficiencies?

Whether AVs will be able to reduce congestion is much less clear. The lesson of the 20th century is that building more roads to ease congestion encourages more car journeys. If robotaxis are cheap and fast, people will want to use them more. Yet there are reasons to think that the roads would become less crowded.

Widespread sharing of vehicles would make much more efficient use of road space; computer-controlled cars can be smart about route planning; and once they are widespread, AVs can travel closer together than existing cars, increasing road capacity.

Yet to think about AVs as a fix for the problems caused by cars is to risk falling into a familiar historical trap. This is exactly how people thought about cars when they first appeared: as a fix for the problems caused by horses. "Cars replaced something that was in many ways far worse," said Donald Shoup of the University of California, Los Angeles. "But because of bad planning, they had unintended consequences."

What might follow from AVs?

Cars transformed retailing, giving rise to suburban malls with lots of shops and plenty of parking. AVs, combined with the rise of e-commerce, could transform it again. "The Walmart of the future might be fleets of vehicles ready to drop off anything that you might get at a Walmart," said Peter Norton of the University of Virginia.

Or you might order an AV to take you home from work, and arrange to have your groceries, or a meal, waiting for you when you climb aboard. And why should shops, restaurants or other facilities be fixed in place? Coffee shops or food stands could restock at a central depot and then migrate to business neighborhoods in the morning and entertainment districts in the evening, suggested Chenoe Hart, an architectural designer at the University of California, Berkeley.

Carmakers are experimenting with delivery vehicles that draw up outside a customer's home, announce their arrival by text message and allow items to be retrieved from a locked compartment by entering a code. Low-cost deliveries using AVs could stimulate local production of all kinds of things, most notably food.

Another possibility, said Johann Jungwirth of Volkswagen, is that restaurants or retailers might cover the cost of travel to encourage customers to visit them. Fancy restaurants might lay on luxury AVs to ferry sozzled customers home, as part of the cost of a meal. Retailers could offer to pay for shoppers' rides. Ride-hailing networks have a lot of customer data that could be used to target in-vehicle advertising. Hail an AV to go to one shop or restaurant, and you might see ads for a rival. Riders may be offered cheaper rides with ads or more expensive ones without them.

Unintended consequences

What unintended consequences might there be? One much-heralded benefit of AVs is that they will offer freedom and independence to people who cannot drive cars: the very old, the very young and the disabled. Such vehicles are already ferrying around people in retirement communities, and one of Google's videos shows a blind man doing errands in an autonomous car.

But AVs could also encroach on freedom by invading people's privacy. And if people no longer drive cars, one consequence may be new forms of segregation, Hart said. In authoritarian societies, AVs could be a powerful tool of social control.

AVs could also trigger a shortage of organ donors (many of whom are young people killed in car accidents) and a drop in smoking (more than half of all tobacco sales in the United States are made at gas stations, which will vanish). And if cars are no longer symbols of independence and self-definition for the young, other things will have to take their place.

Like cars before them, AVs will change the texture of everyday life.