No One at the Wheel by Samuel Schwartz
Self-driving autonomous vehicles (AVs) are inevitably coming. Every major carmaker now has an AV division. An increasing number of jurisdictions are enacting comprehensive AV laws and opening their streets up to AV testing. AVs still lack robustness and are easy to confuse, but they are continuously getting better.
AV makers promise a world where the speed and point-to-point convenience of taxis can be combined with the effortlessness, cheapness, and safety of public transit. They promise to bring easy transportation to people who currently face barriers to access, notably the young, the disabled, the elderly, and the poor. They promise to allow public transit systems to scale up without labor constraints and to become much more efficient than those of today.
Societies will face a messy multi-decade transition in which a growing fraction of cars on the road are AVs. Some usage patterns (e.g. highway trucking, rural public transit, campus or airport shuttles) are much more easy to automate than others (e.g. city driving) and will see mass automation happen sooner. Even when cars are not fully autonomous and remain manually-run, new cars will have far more autonomous features than those of the past. When AVs eventually become dominant, the streets will likely have a blend of public transit, privately-owned vehicles, and private sector fleets that run on a service model, with vehicles from a wide variety of manufacturers. Traffic engineers generally expect manual driving to eventually get outlawed in the US for safety reasons, with their date estimates ranging from 2050 to 2075.
AVs will roll out at different times in different countries and cities. AV adoption will be fastest in places with high labor costs, high-quality road infrastructure, orderly streets, lax data privacy laws, and high public trust in black box technology systems.
The Future of Road Safety
Throughout their short history, traffic accidents have killed 70 million people worldwide. For every person who is killed, eight more are hospitalized and up to a hundred more require outpatient medical care; car accidents have injured at least four billion people in total. 35% of civil lawsuits in the US involve car accidents. Urban planners and traffic engineers have been open-minded and receptive to AVs because they know first-hand how bad the baseline is.
AVs will never be perfect, but removing human error as a cause is expected to reduce accidents by 60 to 90 percent. However, since newer manual cars have more safety features, improvements in crash frequency and severity are expected even if AV adoption is otherwise slow.
Avoiding collisions with pedestrians is a more difficult technical problem for AVs than avoiding collisions with other vehicles. Interactions between cars and pedestrians are heavily driven by hand gestures, eye contact, and other human signals that will be absent in driverless vehicles. On the other hand, if the problem of pedestrian safety is solved, pedestrians may come to expect that AVs will always yield to them and may lose their fear of cars. If pedestrians try to reclaim the streets and jaywalkers create an unbearable experience for cars, carmakers will lobby for anti-pedestrian urban planning measures that are more draconian than anything that exists today.
Avoiding the abuse and weaponization of AVs is still an under-explored topic. All network-connected devices have been cyberattacked and it is inevitable that AVs will be cyberattacked too. Manual cars today are seldom used as battering rams because doing so endangers the driver, but the equation is different when cars can be remotely commanded. AVs will need to be programmed with constraints that they will refuse to break regardless of the instructions they receive, such as a refusal to go onto sidewalks.
The Future of Parking
Parking is an endless source of frustration for urban planners; people like abundant and free parking, but also hate the sort of urban landscape it creates. Car use by households is currently very inefficient, with most cars sitting parked for 95% of the time or more.
In a world where AVs can drop off passengers and then leave on their own, the need for parking spaces on high-value urban land will fall considerably. A large fraction of giant street-grade parking lots and underground parking garages are likely to end up repurposed. There will still be a need for parking, but public transit and private fleet operators will be able to place their depots on low-value land outside city cores.
The Future of Traffic Congestion
Whereas the future looks rosy for parking, urban planners are deeply concerned about traffic. All major studies by traffic engineers expect AVs to increase total vehicle miles driven and for gridlock to get worse in urban cores. The ease and convenience of AVs will prompt people to take trips they otherwise would not have taken, an effect already observed with Uber-style rideshare apps. The effortlessness of AV riding will increase tolerance for longer commutes. Empty cars will spend less of their time on parking lots and more of their time on the streets. Private vehicle owners making short stops will be tempted to tell their cars to locally mope around in circles as a way of avoiding parking costs. Many demographic groups who seldom drive cars or who are unlikely to have licenses (especially the elderly and disabled) are likely to significantly increase their total vehicle miles.
Empty cars causing gridlock will be a new source of public outrage. Cities will have their backs to the wall and will be forced to enact anti-car measures similar to those seen in the world’s biggest cities. Urban core zones will introduce congestion pricing at peak hours, add automated tolling to downtown streets, incentivize private AVs to carry multiple passengers and punish them if they are empty too much of the time, or even disallow all vehicles other than public transit. These new sources of city revenue will recoup the losses due to a reduction in parking fines and traffic tickets.
On the traffic front, AVs will bring a few incremental improvements. A significant fraction of traffic jams are caused by car accidents, which are expected to fall. Another common source of bad traffic flow is cars slowing down when they shouldn’t and causing a ripple behind them; AVs can easily be programmed not to do this. AVs can safely drive next to each other more closely than manual cars, so cities may be able to repaint existing wide freeways and throughfares to add an extra lane. Software will be able to solve the real-time coordination problems that prevent many people from carpooling today. Cars in the same fleet will be able to share real-time information with each other and coordinate their actions to avoid interfering with each other.
Traffic congestion is ultimately a function of the number of vehicles and not the number of passengers. If AVs lead to a culture of widespread public transit use and pay-per-ride models with vehicle sharing, traffic will be manageable. AVs could usher in an unprecedented golden age of public transit, with huge fleets of small buses covering a dense network of routes with a very high arrival frequency. On the other hand, if the old model of widespread car ownership and single-rider commuting continues to prevail, traffic will become worse.
The Future of Cars
Newer cars increasingly resemble computers with wheels, and AVs are poised to continue this trend. Cars will increasingly distinguish themselves with their software and will update it frequently. Cars belonging to pay-per-ride fleets will have intense but short working lives, getting replaced frequently with newer models. As cars become opaque black box systems, DIY or third-party car repair will become increasingly unviable and carmakers will increasingly monopolize the maintenance of their cars.
In a reversal of the long-term trend, the average size of cars may finally start shrinking again in an AV-dominated world. People often buy cars based on the largest usage scenario they need it for, but in a pay-per-ride world, they would be able to hire smaller cars for the majority of scenarios with smaller needs. The improved safety features of AVs may help stop the current arms race in which people buy bigger cars out of fear of being hit by other people’s big cars.
Without the need for manual controls, carmakers will have more freedom in how cars are designed. With passengers having their hands, eyes, and minds freed, car interiors will become increasingly tailored to comfort, entertainment, and productivity. Overnight car trips may become a viable method of long-distance travel, potentially eroding the business of discount hotels and short-distance airplane routes.
The Future of Car Culture
AVs pose an obsolescence threat to millions of workers who currently drive cars and trucks for a living. Since the adoption of AVs is going to be gradual, there will probably never be a sharp employment cliff in which huge numbers of professional drivers suddenly lose their jobs. Professional drivers will likely go the way of manual toll collectors; young people will stop entering the industry (this is already being observed) and existing workers will gradually age out. Automated fleets of AVs will need skilled human workers to operate and maintain them, but the ratio of workers to vehicles is likely to be low.
The unfortunate reality for carmakers is that a world with AVs will not need as many cars as the world currently has. Carmakers will increasingly rely on a fleet service model for their revenues. Liability for AV accidents will rest with carmakers rather than with passengers or owners; consumers would find any other arrangement unacceptable. Legacy carmakers will be tempted to market AVs directly to households for direct ownership, even though a fleet-based pay-per-ride model would be better both for household budgets and for urban traffic. The buy-vs-rent calculation for AVs will be more favorable to renters than it is for manual cars; AV maintenance will be an expensive hassle, newer AV models will be getting deployed frequently, wait times for rented AVs will usually be low, and a higher prevalence of government toll fees will break any illusion that driving is a free action if you own a car outright. A switch from an ownership model to a service model is a cultural shift that is not guaranteed to succeed. However, car culture appears to have peaked with the Baby Boomer generation, and younger generations appear more receptive to the idea of transportation being a service.