Uber and Lyft were supposed to reduce carbon emissions, but they turned out to be polluters. Robotaxis look to repeat some of the same mistakes.
Robotaxi companies are eager to present themselves with a green halo.
“Climate change is the single biggest issue we face as a global community,” Cruise declared in a blog post published on Earth Day 2022. “Each of us has an opportunity to make an impact. Cruise knows the AV industry can –– and should –– help lead the charge.” Its rival Waymo seems to agree. In a post this summer, the company stated, “Cities where we operate gain another zero-emission transportation option, which could help them meet their climate goals.”
Sustainability is critical to the public pitch of robotaxi companies, which are under growing scrutiny after California regulators suspended Cruise’s driverless permit last month due to safety concerns. But the idea that robotaxis will benefit the planet runs counter to what we know about the sustainability of a very similar service: ridehail.
Although climbing into driverless taxis may seem like an entirely novel way to travel, the user experience will be familiar to anyone who has taken an Uber or Lyft: a customer summons a robotaxi using their smartphone, which then picks them up and ferries them to their destination before driving off. It’s basically ridehail, minus the driver.
Robotaxi companies themselves seem to agree; Cruise uses ridehail crash rates as the benchmark for its safety reporting, and Waymo recently unveiled an integration with Uber in the Phoenix area.
Unlike robotaxis, Uber and Lyft have been with us for 15 years — long enough to study and evaluate their impact on sustainability. And based on what we have learned about ridehail, robotaxis are more likely to foul the air than clean it.
Like robotaxi companies today, ridehail executives a decade ago presented themselves as environmental allies. Their core claim was summarized by Logan Green, a co-founder of Lyft, to MIT Technology Review in 2015: “We’re the replacement, the alternative, to car ownership.”
That assertion held intuitive appeal, especially considering ridehail companies’ favorite statistic: the average American car sits unused roughly 95 percent of the time. By making door-to-door trips just a smartphone tap away, Uber and Lyft would empower customers to ditch their cars.
Those liberated from vehicle ownership could then use their ridehail app to access a constellation of climate-friendly mobility modes like transit, bike share programs, and scooters — all of which generate less greenhouse gas emissions per mile than even an electric automobile. The ridehail companies invested directly in micromobility: Uber bought an e-bike share company while Lyft acquired the largest bike share operator in North America.
The companies were bullish about synergies with public transportation. Both Uber and Lyft enabled users to purchase transit tickets on their app, and they offered themselves as a solution to transit’s “first mile / last mile” problem of getting passengers to and from a station. The US Department of Transportation shared their enthusiasm, funding several first mile / last mile pilots involving ridehail companies.
Adding to their environmental bona fides, in 2015, Uber and Lyft unveiled pooled ride products that capitalized on the companies’ prowess in data analysis to match passengers going in the same direction. Those willing to split the trip with strangers would receive a discount and reduce the number of vehicles on the road.
Ridehail executives claimed that their net effect would be fewer cars in a city, spewing air pollution as they go. “Uber can help reduce traffic by taking cars off the road,” vowed Uber executive David Plouffe in 2015. The media helped spread that narrative, with stories like one in The New York Times in 2014, which posited that ridehail could “reduce the environmental toll exacted by privately owned automobiles.”
The reality has been something altogether different.
Research has refuted what used to be ridehail’s most fundamental sustainability argument: that it reduces car ownership and driving. A 2021 study found that car registrations usually rise in a city after ridehail arrives because the number of ridehail drivers who acquire a vehicle exceeds the users who get rid of one.
Other studies have found that Uber and Lyft increase traffic congestion as well as total driving for two primary reasons. First, some ridehail trips would have otherwise occurred on cleaner and more space-efficient modes like biking or transit. Second, ridehail vehicles are often empty because the driver is either cruising streets waiting for the next passenger or en route to pick them up — a phenomenon known as “deadheading.”
According to a 2018 study, even carpool trips cause a net increase in total miles driven. But that finding may be a moot point because practically no one seems to be taking shared rides. Despite massive investments from Uber and Lyft, pooled ridehail has turned out to be a money-losing flop.
Shared trips have intractable downsides, starting with the annoyance of one passenger’s itinerary taking others out of their way. “One of the most compelling reasons to take ridehail is reliability and speed,” said Harry Campbell, founder of The Rideshare Guy blog and podcast. “Shared trips cut into that efficiency.” Pooled ridehail users could also find themselves trapped next to someone unpleasant, without being able to exit the vehicle or switch seats as one could do aboard public transportation.
Lyft has now nixed its pooled ridehail service after briefly trying to restart it post-pandemic. Uber’s offering, meanwhile, is buried within its app. “The proof is in the pudding,” said Campbell. “Pooled ridehail hasn’t worked out.”
Making matters worse for the planet, ridehail has turned out to be more of a competitor than a complement to the greenest transportation modes like transit, bike share programs, and scooters.
“Drivers tend to go where there is a lot of demand, in downtown areas,” Greg Erhardt, a civil engineering professor at the University of Kentucky, told me. “That’s where the mode share of transit, biking, and walking, is relatively high.”
Meanwhile, the much-hyped first mile / last mile (FMLM) connections between ridehail and transit have not materialized. In 2016, Pinellas County, Florida, offered riders $5 off a trip to or from a transit station, but only a few dozen people used it per day, representing less than one transit rider in a thousand. Across the country, a 2022 analysis of the San Francisco Bay Area found that just 0.4 percent of transit riders took ridehail to or from a station.
“The lack of evidence of ride hail working as a FMLM solution is damning,” David King, an urban planning professor at Arizona State, told me in 2019 when I wrote an article in The Drive about first mile / last mile trips. “We don’t see pilots becoming successful and scaling, and we don’t see them leading to increased transit ridership.”
Ridehail’s net effect on public transportation has been devastating. A 2019 study co-authored by Erhardt found that ridehail’s entry into a city typically reduces bus and rail ridership by between 1 and 2 percent per year, compounded annually. The authors concluded that ridehail may be “an important driver of [transit] ridership declines” prior to the pandemic.
Today, ridehail companies have largely abandoned their original vision of fighting climate change by reducing car use. Uber has dumped its e-bike unit, Lyft is reportedly taking offers for its bike share business, and neither company still prioritizes transit scheduling or ticketing within its app.
Instead, the companies have rewritten their sustainability pitch to emphasize commitments to electrify all cars on their platforms by 2030. Although vehicle electrification is a necessary step to combat climate change, EVs still generate greenhouse gasses through their manufacture, charging, and disposal (not to mention air pollution from the erosion of brakes and tires). A 2020 University of Toronto study found that less driving — not just less gas-powered driving — is necessary to prevent a potentially catastrophic two-degree Celsius increase in global temperatures by 2100.
For a while, Uber and Lyft seemed to be champions of the effort to slow climate change by reducing driving. Not anymore.
So what does ridehail’s ignominious sustainability track record portend for robotaxis, assuming that they ultimately scale in the way their executives envision? In short, nothing good.
Robotaxis, like ridehail, are poised to increase total miles driven due to deadheading and transit replacement. And given the similarities between ridehail and robotaxis, the inability of Uber and Lyft to solve transit’s first mile / last mile problem suggests that Waymo and Cruise probably won’t, either. (Data is currently scarce due to the newness of their services.)
Worse, the failure of pooled ridehail suggests that robotaxis, too, will mostly be used for private rides, not those involving multiple parties (something that Cruise and Waymo seem to envision with their new vehicle designs). For some riders — especially women — being stuck inside a driverless vehicle with a stranger may be even less appealing than being in a ridehail car where a driver can at least intervene in an emergency.
Looking to the future, robotaxi companies hope to lower the cost of their technology to the point that trips would be less expensive than in a hailed car with a driver. If that happens, cheap and ubiquitous robotaxis service could damage the planet far more than ridehail has. Many families trade off home size (which is good) and commute time (which is bad) when choosing where to live. If robotaxis make commutes less of a drag, such household calculations would shift toward bigger homes further from the central city. The resulting sprawl would increase pollution not only from additional driving (since destinations in exurban areas are often further away and unreachable by transit or biking) but also from the energy needed to heat, cool, and build larger homes.
As with ridehail companies today, the sustainability claims of Waymo and Cruise focus on their fully electric fleets. “As an all-electric fleet, a trip with Cruise offers every customer an all-electric ride, regardless of their ability to purchase an electric vehicle,” Cruise spokesperson Hannah Lindow wrote in an email. Cruise and Waymo have also committed to using clean energy to provide electricity, but that commitment could be harder to maintain as they expand. And robotaxis require more energy than equivalent driver-operated electric vehicles: a recent MIT study found that autonomous vehicles’ complex computers and sensors could massively increase total power demand.
In all fairness, there are reasons to hope that robotaxis would be at least somewhat less polluting than ridehail. Alex Roy, a co-host of the Autonocast podcast who previously worked with the autonomous vehicle company Argo AI, noted that robotaxi companies, unlike ridehail firms, own their vehicle fleet and are therefore incentivized to minimize unnecessary driving.
“The costs of deadheading in an Uber is absorbed by the driver. In a robotaxi, it’s absorbed by the company,” Roy told me. “In any robotaxi company conversation, we ask, ‘How do we reduce this?’” Theoretically, optimized robotaxi fleet deployments could reduce their total driving compared to ridehail cars owned by gig workers who set their own itineraries and schedules. Cruise’s Lindow emphasized that point: “Through a centrally-managed fleet, AV companies can better reduce consumption to save costs, and position vehicles based on demand, which differs from traditional ridehail.”
Chris Bonelli, a spokesperson for Waymo, wrote in an email that robotaxis would have fewer emissions than ridehail in part because driverless vehicles would be in use almost constantly. “Even assuming human ridehail drivers drive more than the average [ridehail driver], they are still not as efficiently deployed as our fleet,” he said. “Efficiently deployed AVs require less vehicles in total than a distributed system of personally owned vehicles.”
That may well be true, but “better than ridehail” is a low sustainability bar for robotaxis to clear, especially considering the many ways to navigate a city that offer a more promising path toward decarbonizing transportation. Transit, bikes, e-bikes, and scooters (not to mention walking) all produce a fraction of the emissions that an electric car does, regardless of whether it has a driver.
Robotaxi executives seem to realize this, as they often present cleaner modes as members of the same environmentally conscious mobility family. Waymo’s website, for instance, says, “Waymo encourages people to walk, bike, take public transit and when they need a car – have easy access to an electric vehicle” (presumably, an autonomous one).
