Believe it or not, the automobile was once hailed as an “environmental savior.”

In 1900, cities were “drowning” in horse waste. Every day, New York City’s 100,000+ horses produced 60,000 gallons of urine and 2.5 million pounds of manure—which collected on city streets. Manure-borne illness spread by flies killed 20,000 people per year. Likewise, England faced the “The Great Horse Manure Crisis of 1894,” with The Times predicting that, in 50 years, “every street in London will be buried under nine feet of manure.”

The private auto was supposed to solve these problems.

Needless to say, the automobile has not turned out to be an environmental savior, but a far more sinister villain: a chief contributor to climate change, and the cause of major health problems like traffic-related deaths and illnesses from auto emissions.

The challenges with AVs

As electric autonomous vehicles begin to arrive, with promises of ridding cities of congestion and pollution and improving public health, we need to ask what unintended consequences we might be overlooking—so 100 years from now, we don’t rue the very product intended to save us.

For the sake of argument, let’s assume electric autonomous vehicles, or AVs, will be safer and cleaner than their predecessors. What could possibly go wrong?

Well, a number of things when potentially millions of electric AVs are on the roads.

Let’s start with environmental concerns. Lithium and cobalt, two minerals critical to electric AV batteries are toxic, difficult to recycle, and come from countries with lax environmental standards. Lithium extraction requires massive amounts of water (500,000 gallons per ton) and massive amounts of earth to be dug up (1,250 tons per ton of lithium). And cobalt dust, which includes uranium, is inhaled by miners (often children) without proper protection.

Then, there are public health risks. AVs may save lives by reducing accidents; currently, autos kill 1.25 million people each year, with an additional 300,000 premature deaths each year attributed to auto emissions. However, AVs could shorten life spans because people will spend more time sitting in cars and less time biking or walking. (And sedentary behavior contributes to the chronic diseases that kill the most people in the world.)

Traffic could actually get worse. Several studies warn that personal autonomous vehicles could double traffic in already choked cities, because the vehicles would take two trips per day instead of one—taking owners to work and returning home to park, then repeating the same route at the end of the work day. And workers won’t mind long commutes, because they will be able to work, watch TV, have a cocktail, or sleep in cocooned comfort.

In addition, public transit could be gutted. Some cities are already considering postponing investment in public transit, thinking private sector robotaxi services will provide more convenient, cost-effective options. Private operators, however, will have little incentive to reach low-income neighborhoods, further exacerbating income inequality.

There are long-term geopolitical concerns, as well. Cobalt is found almost exclusively in the Democratic Republic of Congo, which has been criticized for exposing workers (including many children) to toxic dust. Lithium is primarily mined in Argentina, Bolivia, and Chile. Other minerals come chiefly from China (graphite) and Russia (nickel). Demand for these minerals has skyrocketed and the mass production of electric vehicles could lead to shortages—and could tilt geopolitical power like the oil crisis did in the 1970s.

Looking for solutions

Fortunately, efforts are underway to avoid some of these scenarios.

The World Economic Forum has launched the Global Battery Alliance to create manufacturing standards to ease recycling, calm demand, and divert spent batteries from landfills, noting, “Standards and regulation are not in place to make repurposing and recycling batteries at the end of their life cost effective.”

Car manufacturers and tech firms are working to combat child labor and human rights in battery mineral mining, with many responding to Amnesty International’s call for an “ethical battery” within five years. And on the public health side, cities including London and Copenhagen are developing “cycling superhighways” to make cycling easier than driving, and incentivizing housing near the routes.

To manage the impact, a study from ETH Zurich recommends that electric AVs should be shared, and that private AVs should be prohibited. Similarly, MIT and The University of Texas found that one shared electric AV could do the work of 10 private vehicles with minimal impact on ride times—potentially reducing traffic by 90 percent.

A shared future

Electric autonomous vehicles have enormous potential to reduce congestion and pollution, save lives, and improve access to transportation for the elderly and disabled. For these benefits to be realized, however, the growing consensus is that these vehicles must be deployed as shared services integrated with high-capacity public transit.

This model—in which the majority of autonomous transit is shared and private vehicle ownership declines dramatically—could prevent many of the possible ills of mass production of AVs. For example, demand for batteries would moderate, traffic would be reduced, and land dedicated to an auto-centric infrastructure could be repurposed, giving people more space to walk and bike in cities. An Australian government report found that in a shared environment, “AVs would roam the city, filling in gaps in the timetables and fixed routes of a superior and cheaper public transport network.” However, for this to happen, “AVs must be shared and not privately owned and must complement a robust public transport system.”

Getting there

MIT and The University of Texas found shared fleets could achieve 10-fold efficiency gains with maximum wait times of five minutes. Considering that it takes 13 to 32 minutes to find parking in global cities, these services could be more convenient than driving.

Achieving the kind of efficiency at scale, though, is not easy. Fleets will need to be carefully orchestrated with algorithmic efficiency far beyond what today’s ridehailing services, for instance, can deliver. Lyft and Uber’s pooling services add 2.6 new miles of travel for every mile of personal autos taken off the road, increasing traffic and frustrating cities and commuters. Uncertainty about excess ride time and wait time has slowed adoption.

There will also need to be some kind of regulation of the number of fleets and vehicles allowed to prowl city streets looking for riders. Market forces may regulate this when fleet owners have to pay for empty vehicle miles—something today’s peer-to-peer ridehailing services don’t have to do. When empty vehicle miles hurt the bottom line, service providers will crave efficiency.

One thing is clear: new types of vehicles alone do not solve environmental or congestion problems. Cities and service providers must consider the nature and quality of services electric autonomous vehicles might enable as well as their social, environmental, and economic impacts—so that, 100 years from now, we aren’t wringing our collective hands over a calamity.