Andrea is a 2017 graduate of Loyola University New Orleans College of Law. While awaiting bar results, she is available for temporary, contract and project-based work and looks forward to helping innovators bring technologies to market as a transactional lawyer.
Autonomous vehicles (commonly known as self-driving cars) are disrupting what it means to get behind the wheel. Cars symbolize freedom on the open road, and provide a great deal of convenience to a licensed driver of any age.[1] However, now that self-driving cars are available, that same freedom and convenience look a bit different. While there is a plethora of possible risks, industries, and laws that will be impacted by autonomous vehicle adoption, this paper narrowly focuses on how the technology will affect land uses (how vehicles, pedestrians and cyclists interact with the built environment). Additionally, this paper also proposes recommendations for policymakers and urban planners.
From Sci-Fi to Neighborhood Streets: Self-Driving Cars are Here with Many Benefits
Autonomous vehicles are no longer a radical notion from a 1980s science fiction movie. In fact, self-driving cars are already on our highways and city streets.[2]The idea sprang from the Defense Advanced Research Projects Agency (DARPA), a division of the United States Department of Defense.[3] The project asked for researchers and innovators to develop “self-driving vehicles capable of driving across the Mojave Desert . . . in a 142-mile course.” [4] Following DARPA’s interest in autonomous vehicle technology, Google set out to commercialize self-driving cars.[5] Today, dozens of technology-based and traditional automotive manufacturers are entering the autonomous vehicle market including Tesla, Bosch, BMW, Intel, Audi and even Uber.[6]
Autonomous vehicles offer many benefits including increased safety, fewer air pollution emissions, and improved mobility for the elderly and disabled.[7] The National Highway Traffic Safety Administration (NHTSA) calculated the following statistics in 2011: 5.3 million crashes, 2.2 million people injured, 32,367 deaths.[8] An earlier study by the NHTSA explains,
. . . most crashes were the result of driver error: about 41 percent were recognition errors (inattention, internal and external distractions); 34 percent were decision errors (driving aggressively or too fast); and 10 percent were performance errors (overcompensation, improper directional control). About 18 percent of the at-fault drivers engaged in at least one non-driving activity; the most frequent was conversation, either with other passengers or on a cell phone. [9]
Similarly, the World Health Organization estimates that “road injury is the worldwide leading cause of death for young people of ages 5 to 29 years old . . . and total traffic related deaths are more than 1,260,000 yearly.”[10] Unlike a human driver, an autonomous vehicle will not be distracted by conversations, can be programmed to prevent overcompensation, will always be scanning for possible problems, and may have faster reaction times.[11]Eliminating the human driver will save lives.[12]
Environmental benefits are possible with the adoption of autonomous vehicles. Fewer cars starting and stopping to the will of an impatient or distracted human driver means less air pollution.[13]Autonomous vehicles are capable of responding to roadway conditions better than fallible humans allowing a shorter driving distance between cars and more vehicles on the roadway for a continuous flow (instead of congestion).[14] If autonomous vehicles opt for hybrid, fuel-efficient, or electric power models, fewer emissions will be generated.[15]
Autonomous vehicles also benefit the elderly and the disabled. An autonomous vehicle that controls driving allows passengers who otherwise would be prevented from operating a vehicle to utilize a self-driving car.[16] This creates greater mobility for the elderly, the visually disabled, and other non-drivers.[17]
Defining An Evolving Species
What exactly is and is not an autonomous vehicle is difficult to describe with so many companies working on the technology. Google describes their technology as capable of “handl[ing] all the driving so you can go from door to door without taking the wheel.”[18] Tesla describes their technology as capable of handling “short and long distance trips with no action required by the person in the driver’s seat.” [19] Is it all driving or only short term driving? Will the answer to that depend on the brand of vehicle purchased?
The NHTSA has better defined autonomous vehicles as part of their efforts in regulating all vehicles and traffic safety by using a classification system.
● Level 0: No automation. A human driver operates the vehicle.[20]An example of this technology is a 1967 station wagon without any special features.
● Level 1: Some automation. A human driver operates the vehicle with some technology-assisting function partly operated by the vehicle’s software system. An example of this technology is adaptive cruise control.[21]
● Level 2: Multi-feature automation. A human driver operates the vehicle with some technology-assisting functions partly operated by the vehicle’s software system specific to primary control functions. An example of this technology is both the lane departure warning and collision avoidance systems operating concurrently.[22]
● Level 3: Limited self-driving automation. A software system operates all of the safety-critical functions of the vehicle with some human control under necessary circumstances. An example of this technology is the current Tesla model X capable of driving on and off the freeway, switching lanes,and parking without driver control.[23]
● Level 4: Full self-driving automation. A software system operates all safety-critical functions, monitors road conditions, and has full operation of the vehicles without human intervention. A human inside the vehicle is not necessary.[24] This is the future.
The Rub: Autonomous Vehicles are Not Part of Urban Design
Autonomous vehicles merging onto roadways disturb existing land uses. [25] For instance, a frequent driving test question might describe the following scenario: two vehicles approach a four-way stop sign simultaneously. Who has the right-of-way? A driver’s education manual might advise both drivers to make eye contact with one driver signaling to the other to be the first to proceed through the intersection. But, what if one of the vehicles is an autonomous vehicle powered by artificial intelligence software? What
happens when the human driver rolls through the stop sign entering the intersection nearly colliding with the autonomous vehicle? What if a pedestrian is in the walkway furthest from the autonomous vehicle? Existing land uses do not yet account for a self-driving car.
The built environment has been planned and developed through the lenses of convenience and human interaction: not through the lenses of 360 degree cameras and an artificial intelligence response system.[26] Autonomous vehicles will disrupt traffic laws, and many of the purposes that current land uses are designed to serve.
Insufficient Traffic Laws
Existing traffic laws do not contemplate a driverless vehicle. Recall the scenario of a human driver and an autonomous vehicle each pulling up to a stop sign simultaneously. If the operator of the level 3 vehicle is asleep or distracted at the stop sign, one of two things could happen. Either, the car safely moves through the intersection by its own accord, or the human driver of the other vehicle is confused about how to proceed risking collision. If the car is a level 4 vehicle at the same intersection there may not be a human inside at all and the human driver of the other vehicle may be unsure of who has the right-of-way. Worse yet, if the stop sign is faded then the autonomous vehicle may not even recognize it and could proceed through the intersection erroneously.
A quick scan of New York’s Vehicle and Traffic statute lists the following topics: signs, signals and markings; right side driving; passing, turns, lane changes and rights-of-way; pedestrian and cyclist interactions; stopping, parking and speeds; types of vehicles including motorcycles; and alcohol and substance prohibitions.[27]
Assuming New York is typical of most state traffic statutes, each of these regulated topics will need amending as autonomous vehicles roam about cities streets with passenger loading zones, and suburbs with school bus crossings. Two necessary revisions are addressed below.
Drivers Versus Passive Passengers
Thus far, traffic laws have presumed a person will be steering, braking, and otherwise operating a vehicle. Now, state statutes must define who constitutes a passenger, a driver, what their respective responsibilities are while in the vehicle, and whether a driver is required in an autonomous vehicle at all times.[28]
The District of Columbia’s autonomous vehicle ordinance defines driver: ““Driver” means a human operator of a motor vehicle with a valid driver’s license.” [29] This definition favors a level 2 or 3 vehicle where a human can supervise the vehicle’s self-driving systems and be able to take back control of steering and braking in case of emergency.[30] A passive passenger is not permitted.
However, assuming that humans will continuously supervise a level 3 vehicle and be ready to take back control is dangerous. [31] Humans are easily distracted: just look at the NHTSA statistics noted above. The “human brain is not good at routine supervision tasks. If a car drives autonomously for many miles without incident, a normal human will no longer pay attention.” [32] Further, because a human and the artificial intelligence software may respond differently to roadway conditions, the human may be unable to distinguish between those situations the car can safely handle and those the human must take back control. [33]
If a pedestrian were to enter a crosswalk without a distracted driver noticing, an accident is likely. By contrast, an autonomous vehicle is not only likely to identify the pedestrian in the crosswalk so that the distracted driver does not have to, but is always sending and receiving data from nearby cameras, a complex wireless communications grid, and artificial intelligence software that has mapped that exact built environment indicating a crosswalk up ahead. Thus, a safer role for humans in autonomous vehicles is passenger, not driver.
California’s proposed regulation for autonomous vehicles defines driver and passenger separately granting both level 3 and 4 vehicles.[34] Drivers are defined as those with operational control when a vehicle is not in “autonomous mode.” [35]Passengers are “occupants” without operational control when the vehicle is in autonomous mode.[36]
The mere contemplation of humans as passengers in a vehicle and traffic statute is progress. This paves the way for a level 4 fleet of vehicles on the roadway helping to realize autonomous technology’s benefits, and at least acknowledges that humans fail at continuous supervision.
Recommendation: Traffic laws must clearly articulate the responsibilities of a driver, and a passenger. [37] Doing so will allow many other traffic considerations, such as responsibility for failing to signal before turning or the appropriate distance needed between cars to safely pass, to be defined with human drivers and autonomous vehicles. When two vehicles simultaneously pull up to a stop sign, a statute can necessitate the natural right-of-way to the human driver and not the autonomous vehicle passenger.
Liability and the Duty of Care
Embedded in traffic law is tort law (and criminal law, but that is not discussed here) imposing a duty of care on all drivers.[38] New York’s statutorily imposed duty of care charges drivers with a duty to maintain a safe speed, reasonably control their vehicles, be alert and responsive to roadway conditions, and use reasonable care to avoid collisions or accidents. [39]
Will this duty be imposed on an autonomous vehicle and manufacturers? [40] Answering this question requires contemplation of how fault should be divided between parties, or whether strict liability is more appropriate and for whom.[41]
Unlike human drivers, an autonomous vehicle is not likely to exceed speed limits, is better able to control the vehicle amongst variable roadway conditions, and is programmed to use care in avoiding collisions. [42]
With multiple real-time cameras, digitally mapped roadways, a series of sensors that detect nearby vehicles and movement, vehicle-to-vehicle communications, and a complex artificial intelligence software system that allows for machine learning, an autonomous vehicle is always on alert.[43] That is, the technology “ receives the data from the sensors, combines it with the map and, using machine learning in a sequential ‘sense – plan –
act’ three step process constantly.[44]
The sensing system will recognize that a ball from a nearby child playing in a park has rolled into the street. After recognizing that an object has entered its upcoming pathway, the vehicle will formulate a plan to either change lanes, release the accelerator, and/or apply the brakes to avoid a collision (while also alerting other vehicles to the obstruction ahead). [45] Then, the vehicle will respond to avoid the collision.
Even with this sophisticated technology, collisions may still result. In May 2016, a Tesla (level 3) plunged into a telephone pole fatally injuring its owner. [46] Immediately before the wreck, the Tesla hit a tractor-trailer, “went under it, [then] emerged on the other side” and traveled directly into the pole.[47] Tesla and to be sure, other autonomous vehicle manufacturers, are actively analyzing what went wrong and why.[48] One possible reason for the collision was “because the crash-avoidance system is designed to engage only when radar and computer vision systems agree.” [49] Designs flaws implicate products liability, typically a strict liability system, leaving manufacturers responsible.
Recommendation: Lawmakers must bring manufacturers to the table when having the liability discussion. [50]Thoughtful compromises between liability and public safety will promote continued industry engagement and wide-scale adoption of autonomous vehicles.
Next, legislation should model statutorily imposed duties of care after Sweden’s insurance regime. Damages for accidents are “automatically compensated by the insurer of the involved vehicles” apportioning liability between all parties and removing the heavy burdens of strict liability on the vehicle manufacturer (because this disincentivizes innovation). [51]
Additionally, regulations could classify the autonomous vehicle as a service instead of a product (implicating strict products liability laws) allowing contract law to apportion risk. [52]
Doing so will allow scenarios like determining responsibility for a collision resulting from failing to signal before turning, or the appropriate distance needed between cars to safely pass, to be better defined. When two vehicles collide at an intersection, a statute can waive manufacturer liability (in some circumstances) and apportion fault equally to all parties should one be an autonomous vehicle.[53]
Stale Land Use Design
Autonomous vehicles are likely to alter how cities are built. For example, if a freeway on-ramp is typically one lane wide with room for five vehicles, the mass adoption of autonomous vehicles (assuming one out of every three vehicles in the on ramp is a self-driving car), means that now six or seven vehicles can fit on the on ramp due to reduced driving distances between cars. Since autonomous vehicles will be constantly scanning conditions and connected to a mass communication grid, the car will know — possibly before human drivers or passengers — that a street has flooded and an alternate route is required. A few of these alterations are discussed below.
Fewer parking lots
Since level 4 vehicles can drive themselves back home, there is no need to park them downtown all day. [54] Better yet, since level 4 vehicles can drive themselves to any destination, car sharing services may prevent the need for all day or long term parking.[55]Prime real estate parking lots will be redeveloped into housing, urban space, etc. [56]
However, reduced parking will negatively impact revenue streams for cities, states, and private parking companies. For example, “San Francisco alone makes an estimate $130 million dollars annually from parking meters.”[57] Additionally, cities generate revenue from parking citation fees. Again in San Francisco, a parking citation fee starts at $74. [58] Since cars will no longer park, and if they do park, will be programmed not to park erroneously, cities will lose a substantial amount of revenue.[59]
Local governments and state transportation departments will lose revenue from parking fines and tolls payments (due to fewer vehicles).[60] Yet, roadway and highway infrastructure costs will increase especially as communications infrastructure capable of handling an increased load is added to each intersection, highway and tollbooth.[61] This will change the taxing scheme, strain funding for infrastructure upgrades, and could have the effect of forcing the autonomous vehicle industry to compensate entirely.[62]
Recommendation: Legislators must acknowledge the economics at play. Both job creation and elimination is possible. Improvements to air quality, and changes in health care costs and insurance rates are likely. Many industries will be impacted and their lobbying efforts will pressure adoption or prevention of autonomous vehicles. New methods for revenue generation may include redevelopment of parking lots to government-owned developments. Further, reasonable taxes on use and ownership of autonomous vehicles can help to bridge funding gaps so that roadway and communications infrastructure upgrades are possible paving the way for a successful autonomous vehicle market. [63]
Lane Changes and Roadway Design
Autonomous vehicles mount cameras allowing for a 360 degree monitoring of roadway conditions, as well as send and receive communications to nearby wireless systems.[64] These features reduce the necessary length of driving distances between cars, the width of lanes, and possibly, the number of lanes.[65]
Roadway engineers commonly face capacity challenges where on and off ramps create bottlenecks, high-occupancy vehicle lanes are timed for commute hours, and limited land is available to expand from a two-lane roadway to a four-lane thoroughfare. [66] Depending on the extent of adoption, autonomous vehicles could assist in adding capacity to roadways.[67]
Recommendation: Legislators and urban planners must engage the autonomous vehicle industry as improvements in land use design are planned, and the necessary upgrades and additions to communication systems are implemented. Land use design must anticipate self-driving cars.[68] Master plans should presume that autonomous vehicles will interact with roadway capacity, redevelopment projects, new roadways and maintenance of existing roadways.[69] Cities and states should create or repurpose lanes to “dedicated autonomous vehicle lanes” much like high-occupancy lanes. [70] As adoption of autonomous vehicles rises, more of the existing lanes can be repurposed for self-driving cars and the increased need for passenger loading areas.
Public Transportation
Cities increasingly face congestion and population growth taxing transportation and transit services. Land use planners frequently struggle with ways to add transit routes to carry more passengers and interconnect to pedestrian and biking pathways. [71] Also, non consumer vehicles such as tractor trailers, haul trucks, buses, and transit trains are all increasingly using autonomous vehicle technology.[72] Soon, haul trucks could be capable of delivering goods driverless, and buses could pickup and drop passengers on non designated routes. [73]
Recommendation: Level 4 vehicles under car share services, in particular, can be utilized as transportation systems without having to fund, design, and maintain new transit routes “provid[ing] the much-needed transit grid.”[74] This can also have positive implications on suburban areas with few or no transit services into cities and more personalized routes. [75]
Conclusion: Policymakers and Urban Planners Must Engage Autonomous Vehicle Manufacturers
Autonomous vehicles are disruptive technology disturbing the ways in which humans interact with the built environment. For each roadway that combines pedestrian, cyclist, and vehicle traffic, an interaction with human drivers and autonomous vehicles will become the new normal. Land use laws and urban planning must adapt. This requires policymakers engage vehicle manufacturers, and embrace technological advancement so that as innovation progresses, the law and public safety is not left behind.
[1]See Edward Humes, The Absurd Primacy of the Automobile in American Life, The
Atlantic (Apr. 12, 2016), https://www.theatlantic.com/business/archive/2016/04/absurd-primacy-of-the-car-in-american-life/476346/
[2]See Jacob D. Walpert, Note, Carpooling Liability?: Applying Tort Law Principles to theJoint Emergence of Self-Driving Automobiles and Transportation
Network Companies, 85 FORDHAM L. REV. 1863, 1870 (2017) (“In Fall 2016, Uber, a TNC [transportation network companies or ridesharing services] and the world’s most valuable start-up began deploying self-driving vehicles in Pittsburgh, Pennsylvania. Several months later, Uber expanded its pilot program to Arizona . . .”).
[3]See id. at 1867.
[4]Id.
[5]See id. at 1867-1868 (“[T]he winner of the second DARPA challenge was a Stanford University team led by Sebastian Thrun, a Google engineer and coinventor of the company’s “Street View” mapping service. Along with Google cofounder Larry Page, Thrun and his engineering team were early promoters of self-driving vehicles’ potential to lower energy costs and make highways safer.”).
[6]See Christina Mercer, 11 Companies Working On Driverless Cars,
Techworld (Mar. 14, 2017), http://www.techworld.com/picture-gallery/big-data/-companies-working-on-driverless-cars-3641537/
[7]See Todd Litman, Autonomous Vehicle Implementation Predictions, Victoria Transp. Policy Inst. 10 (May 1, 2017) http://www.vtpi.org/avip.pdf
[8]
Christopher B. Dolan, Self-Driving Cars and the Bumpy Road Ahead, Am. Assoc. for
Justice (Feb. 2016) https://www.justice.org/what-we-do/enhance-practice-law/publications/trial-magazine/self-driving-cars-and-bumpy-road-ahead
[9]Id.
[10]Mahmood Hisham Shubbak, Self-Driving Cars; Legal, Social, and Ethical Aspects,
(Dec. 18, 2013) https://ssrn.com/abstract=2931847
[11]See id.
[12]See Shubbak, supra note 10.
[13]See Litman, supra note 7.
[14]See Litman, supra note 7; and see Shubbak, supra note 10.
[15]See Litman, supra note 7; and see Shubbak, supra note 10.
[16]See Litman, supra note 7; and see Jerome M. Lutin ET AL, The Revolutionary Development of Self-Driving Vehicles and Implications for the Transportation Engineering Profession , ITE Journal 30-31(July 2013) http://bit.ly/2qj8KqB
[17]See id.
[18]Waymo, https://waymo.com/tech/ (last visited May 8, 2017).
[19]Tesla, https://www.tesla.com/autopilot (last visited May 8, 2017).
[20]
See Dolan, supra note 8.
[21]See Dolan, supra note 8; Shubbak, supra note 10.
[22]See id.
[23]See Dolan, supra note 8; and see Tesla, supra note 19.
[24]See Dolan, supra note 8; and see Shubbak, supra note 10.
[25]See Nat’l Highway Traffic Safety Admin., Federal Automated Vehicles Policy,
https://one.nhtsa.gov/nhtsa/av/av-policy.html (last visited May 8,
2017).
[26]See Sam Lubell, Here’s How Self-Driving Cars Will Transform Your City,
WIRED.COM (Oct. 21, 2016),
https://www.wired.com/2016/10/heres-self-driving-cars-will-transform-city/
[27]See 7 N.Y. Veh. & Traf. Law § 23-34C (West, Westlaw through
L.2017).
[28]See Emma Wright, The Legal Challenges for Autonomous Vehicles,
BETANEWS.COM, (Nov 2016), https://betanews.com/2016/10/21/legal-challenges-self-driving-cars
[29]D.C. Code Ann. § 50-2351–50-2354 (2001).
[30]See Alexander Hars, Supervising Autonomous Cars on Autopilot: A Hazardous Idea, Driverless-Future.Com (Sept. 16, 2013), http://www.driverless-future.com/?cat=9; Wright, supra note 28.
[31]See Hars, supra note 30.
[32]Id.
[33]See id.
[34]See Cal. Dep’t. Of Motor Vehicles, Proposed Driverless Testing and Deployment Regulations (Mar. 10, 2017), https://www.dmv.ca.gov/portal/dmv/detail/vr/autonomous/auto
[35]Id.
[36]Id.
[37]See Nat’l Highway Traffic Safety Admin, supra note 25.
[38]See Walpert, supra note 2 at 1878.
[39]See id. at 1888-1889.
[40]See id.; and see Andrew P. Garza, “Look Ma, No Hands!”: Wrinkles and Wrecks in the Age of Autonomous Vehicles, 46 NEW ENG. L. REV. 581, 591-609 (2012).
[41]See Walpert, supra note 2 at 1878; and see Garza, supra note 40.
[42]See Richard Kemp, Legal Aspects of Artificial Intelligence, http://www.kempitlaw.com/legal-aspects-of-artificial-intelligence/ (last visited May 8, 2017).
[43]See id.
[44]Id.; and see Tao Jiang ET AL, Self-Driving Cars: Disruptive or Incremental?, Applied
Innovation Review (June 2015), http://cet.berkeley.edu/wp-content/uploads/Self-Driving-Cars.pdf
[45]See Jiang, supra note 44.
[46]See Alexander Hars, Fatal Tesla Accident Exposes Fundamental Flaws in the Levels of Driving Automation Framework, Driverless-Future.Com (July 11, 2016), http://www.driverless-future.com/?p=955
[47]Id.
[48]See Anjali Singhvi & Karl Russell, Inside the Self-Driving Tesla Fatal Accident, N.Y. Times(July 12, 2016) http://nyti.ms/2pejh6k
[49]Id.
[50]See Nat’l Highway Traffic Safety Admin, supra note 25.
[51]See Dolan, supra note 8.
[52]See id.
[53]See Shubbak, supra note 10 (describing that legislation should provide incentives for manufacturers to design software that can not be hacked to adjust the vehicle’s programed safety responses or share private information).
[54]See Chris Giarratana, Are American Cities Ready for Self-Driving Cars?, Safety Resources Center (Dec. 8, 2016), https://www.trafficsafetystore.com/blog/self-driving-cars-impact-urban-planning/
[55]See id.; and see Eric Jaffe, Why Aren’t Urban Planners Ready for Driverless Cars?,
CITYLAB.COM (Dec. 8, 2015), http://www.citylab.com/cityfixer/2015/12/why-arent-urban-planners-ready-for-driverless-cars/419346/
[56]See Lubell, supra note 26.
[57]See Giarratana, supra note 54.
[58]See id.
[59]See id.
[60]See id.
[61]See id.
[62]See id.
[63]See id.
[64]See Litman, supra note 7.
[65]See Jaffe, supra note 55; and see Lubell, supra note 26 (“For land use design, “extra lanes are unnecessary in an age of driverless cars, which can safely operate closer together and thus serves as a de facto road expansion by themselves.”).
[66]See Jane Bierstedt ET AL, Effects of Next Generation Vehicles on Travel Demand and
Highway Capacity, FP Think 19-25 (Jan. 2014), http://orfe.princeton.edu/~alaink/Papers/FP_NextGenVehicleWhitePaper012414.pdf
[67]See id.
[68]See Alexander Hars, Five Guiding Principles for Autonomous Vehicle Policy,
Driverless-Future.Com (Oct. 20, 2014), http://www.driverless-future.com/?p=683
[69]See Jaffe, supra note 55.
[70]See Litman, supra note 7.
[71]See Giarratana, supra note 54.
[72]See Erick Guerra, When Autonomous Cars Take to the Road, Am. Planning Assoc. (May 2015), https://www.planning.org/planning/2015/may/autonomouscars.htm
[73]See Lutin, supra note 16.
[74]Giarratana, supra note 54.
[75]See Guerra, supra note 72, and see Jaffe, supra note 55.