Factsheets

Autonomous vehicles (AVs) use technology to partially or entirely replace a human driver in navigating vehicles, responding to traffic conditions, and avoiding road hazards.¹ The U.S. National Highway Traffic Safety Administration (NHTSA) uses a classification system developed by the Society of Automotive Engineers (SAE), based on the degree of human intervention.²

Development of Autonomous Vehicles

• In 1980, university researchers began work on two AV concepts, the first required automated roadways, while the other did not.¹
• From 2003–2007, The U.S. Defense Advanced Research Projects Agency (DARPA) held three grand performance challenges with up to $2M incentives to accelerate advancements in AVs.¹
• No vehicle successfully finished the 2004 challenge—the best covered just 8 mi. In 2005, five completed all 150 mi. In 2007, six teams finished the third challenge; a 60-mi course navigating a dense urban area while obeying traffic laws.¹
• In 2015, the University of Michigan built Mcity, the first testing facility for research on the safety, efficiency, accessibility, and commercial viability of AVs.⁴
• Unmanned aircraft systems (UAS), or drones, are now deployed for commercial uses (like last-mile package delivery), medical supply transportation, and inspection of critical infrastructure.⁵

Autonomous Vehicle Technologies

AVs use a combination of sensors to detect their surroundings, including obstacles and signage, as well as identifying appropriate navigation paths. Artificial intelligence software in the vehicle uses sensors, cameras, and satellite maps to simulate human perception and decision-making, and to anticipate landmarks, traffic lights, and signs.⁶

Autonomous Vehicle Technologies^1,7,8,9

Image

Current and Projected Market

Key Market Leaders

• In 2021, industry leaders believed North America would be the first to deploy a Level 4 AV. By 2023, predictions were evenly split between North America and China.¹⁰
• Waymo’s ride-hailing service has driven over 71M mi on public roads without a human driver and billions more in simulation.^11,19 In 2025, they launched a partnership with Toyota to develop a next-generation personal use vehicle.²⁰
• Teslas have driven 3B mi in Autopilot since 2014,¹² reporting 85% less accidents than the U.S. human driver average.³⁴
• Other major players include Audi, BMW, Daimler, GM, Nissan, Volvo, Bosch, Continental, Mobileye, Valeo, Velodyne, Nvidia, Ford, and many other OEMs and technology companies.^6,13

Regulations, Liability, and Projected Timeline

• Regulation will impact AV adoption. The U.S. has no national standards for AVs, allowing states to determine their own.¹⁴
• In 2018, Congress introduced the AV START Act to establish a framework for testing, regulating, and deploying autonomous vehicles. The legislation ultimately failed to pass both houses.¹⁵
• By February 2020, 29 states and D.C. had enacted legislation regarding the definition, use, and liability of AVs.¹⁶
• In 2025, the U.S. Department of Transportation unveiled a new framework to modernize federal standards for commercial AVs, including new safety reporting guidelines and deregulation to accelerate development.
• The 2016 Tesla Model S fatality highlighted the need for clear liability laws in AV crashes. Liability depends on various factors, including whether the vehicle is operated appropriately to its level of automation.^17,18
• Level 5 availability before 2035 is unlikely.³⁶

Limitations and Barriers

• Advancing vehicle autonomy becomes increasingly complex and costly at each step. Level 2 systems are becoming standard, but require the driver’s constant attention. Current driver attention sensors are easily defeated, requiring more secure solutions like eye-tracking.³⁶
• Level 3 autonomy poses a major liability shift from driver to automaker, slowing industry adoption. Level 4 AVs (e.g. robo-taxis) are limited to specific geofenced areas and conditions, and consumer demand for personal vehicles is too low to justify costs. Without strong willingness to pay for Level 4 capabilities, mass-market deployment is unlikely in the next decade.³⁶
• Additionally, data security concerns, vulnerability to cyberattacks, and public distrust impede the adoption of AVs.⁶

Impacts and Solutions

• AVs alone are unlikely to have direct impacts on energy use and GHG emissions. Combined with other technologies and new transportation models, they could have significant economic, environmental, and social benefits.^21,22
• With eco-driving, intersection connectivity, and faster highways, energy use and GHG emissions can reduce by 9%.²³

Metrics and Associated Impacts

• Congestion is predicted to decrease, reducing fuel consumption by 0-4%. Decreased congestion is likely to lead to increased vehicle-miles traveled (VMT), partially offsetting these gains.²¹
• Eco-Driving, a set of practices that reduce fuel consumption, is predicted to reduce energy consumption by up to 20%.21 If AV algorithms do not prioritize efficiency, fuel efficiency may actually decrease.²⁴
• Performance, such as acceleration, is likely to be de-emphasized as comfort and productivity become travel priorities, leading to a 5-23% reduction in fuel consumption.²¹
• Improved Crash Avoidance, due to the increased safety features of AVs, may allow for the reduction of vehicle weight and size, decreasing fuel consumption 5-23%.²¹
• Vehicle Right-Sizing, the ability to match the utility of a vehicle to a given need, could decrease energy consumption 21-45%, though the full benefits are only likely when paired with a ride-sharing on-demand model.²¹
• Higher Highway Speeds are likely due to improved safety, increasing fuel consumption 7-30%.^21,25
• Travel Cost Reduction, due to decreased insurance cost and improvements in productivity and driving comfort, could result in increased travel, increasing energy consumption 4-60%.²¹
• New User Groups could increase VMT and fuel use by 2-10%.²¹
• Changed Mobility Services, such as an increase in ride-sharing could reduce energy consumption 0-20%.^21,26

An accurate assessment of these interconnected impacts cannot currently be made. One optimistic scenario projected a 40% decrease in road transport energy, while the most pessimistic scenario projected a 105% increase.²¹

Potential Benefits and Costs

• 42,795 people died in vehicle crashes in 2022.²⁷ 94% of crashes are due to human error. AVs have the potential to eliminate human error and decrease deaths.²⁸
• Compared to the average human driver, Waymo had 88% fewer serious injury crashes and 93% fewer involving pedestrians.¹⁹ AVs crash reduction potential could save $190B/yr.²⁹
• The U.S. AV market is expected to grow to over $75B in 2030, an increase of 350% from 2023.³⁰
• The last-mile AV energy savings for public transportation were over 33% compared to private vehicles.³¹
• Potential benefits include improved safety, quality of life, public health, mobility, and accessibility,—especially for the elderly and disabled. ^1,14,32
• AVs may also reduce energy use, environmental impacts, congestion, transportation costs, and promote greater adoption of car sharing.^14,33
• Potential costs of AVs include increased congestion, vehicle-miles traveled (VMT), urban sprawl, and total travel time.^25,34 High upfront costs raise social equity concerns.¹
• Additional challenges stem from heightened concerns about security, safety, and public health,25,33 and affects on other transport modes.¹⁴