Autonomous vehicles hold great promise to remake our society and economy, but not everyone is along for the ride. There are, of course, the concerns about industries that self-driving cars will disrupt and the people they will put out of work. Skeptics like Fox News pundit Tucker Carlson have suggested we might be better off if the government were to stop this transition from happening at all, as “nothing is more American than driving.”

The evidence actually points in the opposite direction. Self-driving cars may be so much safer — eliminating most of the 94 percent of motor-vehicle accidents that are due to driver error — that it won’t be long until we have to ponder the previously unthinkable: banning human-piloted vehicles from the road.

While we still are years away from large-scale deployment, self-driving cars are making their way into select cities and testing centers throughout the United States, as state and federal policymakers experiment with how to regulate them. In addition to saving lives and preventing crippling injuries, the adoption of autonomous vehicles promises hundreds of billions in savings on fuel, reduced property damage and increased productivity during our daily commutes.

But today’s human-operated vehicles soon will be more than just obsolete. Because of their inability to integrate with traffic safely, they’ll be a hazard to others on the road. Automation proponent Elon Musk is among those now arguing we may soon “outlaw driving cars because it’s too dangerous.

At the very least, it will almost certainly be necessary to prohibit them from certain areas, just as we don’t let people ride horses on a major freeway. At the end of the day, imposing some new restrictions on human-operated vehicles may simply be unavoidable.

It’s not hard to imagine the benefits of having areas dedicated exclusively to automated vehicles. Removing human error from the equation would extend safety benefits to everyone. Traffic efficiency would improve, as automated vehicles coordinate their operations, either through vehicle-to-vehicle or vehicle-to-infrastructure communication systems, or through decentralized apps.

Our commutes, which currently average 25 minutes each way, will become shorter and less stressful. Most analysts expect the self-driving future will be one in which fewer people will own cars or need parking spaces for those cars, as fleets of ridesharing services fill our commuting needs. That means our infrastructure needs would diminish massively.

The human-operated cars of the future could be an anachronism that we simply have to accommodate, like Amish horse-drawn carriages. But it’s probably not slated to be an activity of the past. Just like horses, which are still legal on many American roads, there are many reasons human-operated cars won’t be banned outright.

The political costs of an outright ban would be prohibitive, at least over the near term. Americans also use their cars for more than just commuting. We buy muscle cars to gun for potential glory between traffic signals, and sport utility vehicles so we can take spontaneous off-road weekend adventures.

Moreover, spurred by ever-improving reliability, the average car on America’s roads is now more than 11 years old. The length of financing periods for vehicles also continues to grow longer; last year, it was nearly 70 months. That means Americans who buy cars today are making long-term commitments — both to use them for a long time and to be paying for them for a long time.

An outright ban also may not be necessary. Even Amish buggies now include many safety features that weren’t present in their original designs, including modern brakes, lights and electrical systems. In the same way, many of the technologies that promote safety in autonomous vehicles can be — and already are being — applied to human-operated cars.

A study by the Insurance Institute for Highway Safety has found these collision-avoidance systems to be effective at reducing the number and severity of accidents, estimating that current technology can prevent one out of three fatal crashes. As these systems get better and sophisticated sensors become more ubiquitous, they will rapidly approach the safety of fully autonomous vehicles.

But we also should expect to see the creation of zones in which traditionally operated vehicles are restricted, perhaps concentrated in urban areas with heavy traffic like San Francisco or Manhattan. As a first step, traffic planners could look to tackle high-congestion zones by assessing tolls on human-operated vehicles, from which automated vehicles would be exempt. Over time, such zones would expand to roads limited to self-driving vehicles only.

It might even make sense to establish “park and ride” arrangements similar to those already popular today, where drivers would leave their human-piloted cars and hop in a self-driving vehicle.

With the labor cost of drivers largely factored out of the mix, ridesharing services would become much cheaper. It might also be the case that traditionally operated vehicles outfitted with self-driving systems would be able to integrate with “automated only” zones, in the way that Tesla owners can engage the autopilot function.

Regardless of how limits are applied, there will continue to be places for traditionally operated vehicles to operate for the foreseeable future. Be it driving through poorly mapped areas, off-road excursions or driving purely for pleasure, human-operated vehicles will persist, despite their shortcomings in safety, efficiency and environmental impact.

Automated vehicles don’t have to mean the end of the golden age of driving in America. Driven vehicles might even experience greater freedom once they are relieved of the presence of thousands of other inattentively operated human vehicles, not to mention all the congestion that comes with them. In exchange for giving up certain areas of exclusive automated operation, the rest of the roads will open up for those with an intense preference to drive. That’s not a bad trade.