Equity issues associated with the widespread implementation of autonomous vehicles

Abstract

Autonomous vehicles (AVs), either shared or privately owned, are predicted to become a common transport mode used by the general population in coming decades. Policies governing the use of AVs may increase or decrease social inequities. This review synthesizes existing literature and provides policy recommendations to enhance equity as the use of AVs becomes more widespread. We identified nine areas in which AVs could impact equity: (i) assessment of community mobility needs and priorities, (ii) education and outreach, (iii) disparities in infrastructure quality, (iv) equitable distribution of customer services, (v) access to AVs by persons with low incomes, (vi) shared infrastructure services, (vii) barriers to shared AV use, (viii) access to AVs by persons with disabilities and (ix) disruption of existing transportation jobs. Recommendations for promoting equitable use of AVs include (i) policies governing how jurisdictions oversee AV implementation and (ii) policies addressing how jurisdictions issue permits to AV service providers. Oversight policies include ensuring input from disadvantaged communities, providing subsidies for low-income users, establishing ride-sharing rules to protect vulnerable populations, reviewing the equity implications of proposed AV infrastructure improvements, providing retraining opportunities for those who may lose jobs due to AV implementation and monitoring the impact of AV policies implemented. Permitting processes include ensuring equitable access to AVs for low-income, minority, and older users and persons with disabilities, ensuring equitable distribution of AV service areas and verifying that data from all communities are incorporated into the artificial intelligence algorithms used to guide AVs.

autonomous vehicles (AVs), equity, low-income, disabilities, shared autonomous vehicles, shared AVs

INTRODUCTION

Access to jobs, housing, education, food, healthcare and social relations are necessities for all individuals and communities. These necessities can be met through a combination of mobility, proximity (e.g. mixed use neighborhoods) and technology (e.g. teleworking)—lack of which leads to exclusion from economic opportunities and participation in society [1, 2]. Over the past half century, the primary mode of mobility for most trips in the USA has been by automobile, leading to negative consequences on our environment, public health, and the distribution of employment opportunities across geographies and social groups [2–4]. Transportation policies in the USA have historically been designed to favor automobile users [2, 5]. The transportation sector in the USA also has a history of injustices against disenfranchised communities including community severance by multilane highways and poor public transit infrastructure [4, 6–9]. Forty-five percent of the US population lacks access to public transportation [10–12]. Groups that may have limited access to transportation include low-income people, racial minorities, immigrants, women, people with disabilities, seniors, teenagers and rural residents [6–9, 13, 14].

An autonomous vehicle (AV) is a vehicle capable of driving itself ‘from a starting point to a predetermined destination in “autopilot” mode using various in-vehicle technologies and sensors, including adaptive cruise control, active steering (steer by wire), anti-lock braking systems (brake by wire), GPS navigation technology, lasers and radar’ [15]. Although the timetable is uncertain, widespread use of AVs is likely in coming decades.

AVs have the potential to reduce inequities in transportation if their implementation is designed to ensure equitable and affordable distribution of the mobility services they provide. Equity implies fair inclusion, such that access and participation in offered services is available to all persons, regardless of income, race, religion, age, gender, disability or other socio-economic factors—to a degree that fulfills basic transportation needs with reasonable cost, time, safety and comfort. To attain social equity when planning for AV infrastructure and services, disadvantaged communities may require greater investment and active involvement of affected residents [7, 16, 17]. Policymakers need to evaluate future widespread AV technology and use for technical design, manufacturing, and distribution, and for policy aspects such as incentives for shared riders and workforce disruption [2, 18, 19]. Widespread adoption of AVs may alter labor demand, private vehicle ownership and urban land use. While AV implementation has potential to offer increased safety, access and travel efficiencies, it may also lead to negative impacts such as increased vehicle miles traveled and urban sprawl, exacerbating pollution and socioeconomic inequality. The resulting externalities are likely to be inequitably distributed. Policies to ensure equitable access to AVs, such as subsidies, may increase these externalities and inequities [1, 18, 20, 21].

This article reviews AV literature, focusing on electric AVs for individuals or small groups with Level 4 or 5 automation [22]. While primarily focused on the USA, results may be applicable to the global transportation community. Equitable AV policies can support low-income individuals, people of color and those with mobility challenges [18, 23]. Policies should ensure equitable AV service provision across all geographic locations [2, 7, 23, 24]. While widespread AV use may not be the best way to reduce inequities in accessibility and transport externalities, our intent is to identify considerations, challenges and recommendations to maximize benefits and minimize harms of widespread AV use.

METHODS

This review arose from discussions with the Health and Equity Subcommittee of the Washington State Transportation Commission Autonomous Vehicle Work Group (https://wstc.wa.gov/autonomous-vehicle-work-group/). This group was mandated to identify areas of concern related to the future use of AVs in Washington state and to recommend policies and laws to address those concerns. In addition to health and equity, this Work Group focused on AV issues such as liability, licensing, safety, systems technology, data security, infrastructure and workforce.

We initially identified equity concerns in seven papers that reviewed the potential impacts of AVs on health [24–30]. Next, we used a snowballing technique to identify other relevant papers that cited these papers or were cited by these papers. Keywords including ‘autonomous vehicles’, ‘equity’, ‘low income’, ‘disability’, ‘women’, ‘senior citizens’ and their synonyms were used to search Google Scholar and PubMed. Our review included over 60 academic papers, published and unpublished documents, websites of research organizations and policy memos from government and private organizations. We organized the issues identified into nine areas in which AVs could positively or adversely impact public health and equity, based on their subjective importance to equity, the frequency of their mention and their potential usefulness in informing policy recommendations. We explored policy approaches that could be used to address each equity issue. This was not a systematic or scoping review but was done with a specific lens for policy development.

RESULTS

In our synthesis of reviewed literature, four areas in which AVs could impact equity relate to the initial development and deployment of AVs: (i) assessment of community mobility needs and priorities, (ii) education and outreach, (iii) disparities in infrastructure quality and (iv) equitable provision of customer services. Another four areas relate to equitable access to AVs: (v) access by persons with low incomes and/or lacking access to smartphones and credit cards, (vi) shared infrastructure services, (vii) barriers to shared AV use, including women’s safety, and (viii) access by persons with disabilities. The final area relates to long-term impacts: (ix): the potential disruption of existing transportation-related jobs. These factors operate on the community and/or individual levels [31]. The following sections describe each topic’s significance, cite relevant literature and suggest policy recommendations (Table 1).

Table 1

Equity topics impacted by AVs, importance and possible policy approaches

Equity topic of concern	Importance	Possible policy approaches
1. Assessment of community mobility needs and priorities (References:  7, 18, 25, 30, 32, 33)	Community needs should be assessed to tailor AV services to varied geographic and social requirements. Community priorities may include improvement of existing public transit and sidewalks before improvements to facilitate AV implementation.	Conduct community mobility needs assessments including active outreach to representatives of disadvantaged populations as part of AV policy and planning development processes. Promote fleets of shared AVs rather than individually owned AVs to enhance community benefits and reduce inequities. Use tactical urbanism (low-cost temporary changes to the built environment) and research studies to envision equitable scenarios and goals.
2. Education and outreach (References:  2, 4, 7, 13, 25, 34)	Persons with low education or limited English proficiency may be less aware of AV issues and less able to provide informed input in community participatory processes.	Increase public awareness by conducting active educational outreach about AVs to persons with low education before gathering community opinions Provide outreach in multiple languages to assist community residents who have limited English proficiency
3. Disparities in infrastructure quality (Reference:  14)	Improvements in existing roads to accommodate AVs may be disproportionately made in higher income areas. Ensuring equitable investment in infrastructure improvements facilitates equitable provision of AV services.	Improve lane markings and signals and overall road maintenance to benefit all road users, through public and private funding sources. Review types and distribution of proposed roadway improvements to ensure equitable distribution.
4. Equitable access to AV services across all neighborhoods (References:  2, 7, 19, 26–28, 35, 36)	Ensuring access to AV services for persons in neighborhoods with racial and ethnic minorities would help reduce inequalities in existing transportation systems Optimize AV testing in all typologies of geo-spatial regions with expected AV service to promote equity.	Review proposed AV geographic service areas to ensure equity in access across all neighborhoods. Ensure that data used for AI algorithm training are collected equitably from minority populations rather than predominantly from White populations. Update laws and policies to cover liability and insurance related to crashes involving AVs.
5. Access to AVs by persons with low incomes (References:  2, 20, 27, 28, 37–39)	Ensuring affordable access to AV services for persons with low incomes would help reduce inequities in existing transportation systems Low-income persons may not have smartphones or credit cards needed to use AVs. Seniors may not have smartphones. Access to internet and quality of internet service impacts communication with AV services.	Provide targeted subsidies to support AV use by low-income persons Engage persons from disadvantaged communities in development of AV policies Incentivize the implementation of affordable shared AVs Offer alternative communication and payment options for low-income persons Offer simpler smartphones to seniors Improve access, quality, and affordability of internet across various social groups
6. Shared infrastructure services (References:  27, 40)	Sharing of infrastructure and services between AVs and other transport modes helps lower overall investment and maintenance costs. Infrastructure investments are more equitable if they benefit multiple transport modes.	Prioritize investments in filling potholes, improving lane markings and providing pedestrian and bicycle infrastructure to help AVs and other transport modes Give lower priority to roadway-to-vehicle communication technologies, detailed mapping, and designated lanes that primarily help AVs
7. Barriers to shared AV use: safety for women and other vulnerable groups (References:  7–9, 41–47)	Women may be more concerned than men about sharing an AV ride with a stranger. Many vulnerable persons based on age, disability, gender, race or ethnicity may be less comfortable sharing an AV with a stranger.	Develop capacity for vulnerable persons to select AV ride sharing partners without creating discrimination concerns Include camera surveillance and voice-activated emergency call buttons inside AVs Provide option for users to automatically notify emergency contact whenever riding in an AV
8. Access to AVs by persons with disabilities and seniors (References:  7, 24, 25, 28, 30, 48–54)	AVs have the potential to reduce inequities in access to transport mobility, increase personal independence and reduce social isolation for persons with disabilities and for seniors who may have motor, sensory and cognitive limitations.	Provide AV services that exceed minimum ADA requirements for accommodating mobility, vision, hearing and cognitive disabilities Require a percent of AVs to accommodate needs of wheelchair users Provide targeted subsidies and assistance for persons with disabilities and seniors Support research to understand factors affecting acceptance of AVs as a transportation choice
9. Potential disruption of existing jobs in transportation (References:  24, 27, 55–57)	Once AVs are in widespread use, there may be fewer jobs related to driving taxis, buses and trucks which would disproportionately impact persons with low levels of education.	Develop job retraining targeted to transportation-related workers whose jobs may become obsolete; new jobs are likely to become available in AV maintenance and dispatching

Assessment of community mobility needs and priorities

The private sector may focus their efforts to implement AVs in moderate to high income neighborhoods rather than disadvantaged communities. To promote equity, public policies guiding the implementation of AVs need to address the mobility needs and priorities of persons with low incomes and persons with disabilities.

A community mobility needs assessment can identify appropriate transportation modes for specific needs and geographic context [18]. Transportation planning agencies should invest in active outreach in hard-to-reach communities to devise policies that meet community needs and decrease unintended, adverse outcomes for vulnerable communities. Outreach should consider future scenarios for widespread AV use and discuss alternatives, as low-income individuals may prefer investments in public transit. AV considerations may be included in the US Department of Transportation’s Coordinated Public Transit Human Services Transportation Plans [58]. These plans are developed through the participation of seniors, individuals with disabilities, low-income persons, and representatives of communities and organizations to guide transportation funding priorities.

The integration of AVs into existing transportation systems determines their beneficial or adverse impacts. Key factors include ownership models, fuel sources, policy frameworks, government regulations and user experiences [25]. A promising approach involving fleets of electric AVs, known as FAVES, offering various societal benefits [18]. These fleets can be maintained with lower per-ride costs compared to individually owned vehicles. Electric vehicles generally emit fewer pollutants than those using fossil fuels. Individually owned AVs with fossil fuel dependence have adverse effects on health, the environment and transportation equity [25]. Hawaii exemplifies an equity-promoting strategy, incorporating AVs into a policy for accessible, automated, connected, electric, and shared (A²CES) mobility to foster environmental health and economic opportunities [32].

Education and outreach

Public acceptance of AVs is linked to awareness, safety perceptions and perceived benefits. Insufficient understanding hinders effective participation in policy discussions. For instance, a 2017 London study found only 6% claimed knowledge of AVs [34]. In Australia, 21% were aware of potential crash reduction benefits, and 60% in La Rochelle, France, expressed willingness to use AVs [59, 60]. A 2022 US survey noted increased enthusiasm post-COVID, with generational divides; ‘Gen-Z-ers’ and ‘Millennials’ displayed higher trust in AV taxis than ‘Boomers.’ Safety, comfort, solitude and customizability mattered to respondents. Forty-three percent felt human drivers are safer, and 33% considered AV ride-hail services safe [61, 62]. Outreach to disadvantaged communities is vital to address concerns about affordability and infrastructure access [3]. A UK survey with 211 participants with visual impairment highlighted expectations for AVs: mobility independence, skepticism, safety and affordability [63].

Encouraging positive AV reception involves financial subsidies, education, outreach and awareness [4]. Overall awareness and readiness for AVs remain low; trust can be developed with exposure and usage frequency [4, 25]. A US survey revealed increased comfort with AVs after riding, with 53% respondents feeling comfortable with safety features [61, 62]. The GATEway Project in England engaged the public through driverless pod rides, revealing 79% overall positive views - 100% positivity among mobility-challenged participants, 53% emphasizing environmental benefits and 39% highlighting safety for pedestrians and cyclists [64]. Demonstration projects and community-engaged research in public spaces, especially targeting disadvantaged populations with limited AV knowledge, can enhance public acceptance. Outreach programs should offer information in multiple languages.

Disparities in infrastructure quality

Various infrastructure improvements may help facilitate the widespread use of AVs. Some improvements may favor access to improved transportation for all populations and therefore support equity. For example, reducing the complexity of a multimodal transportation environment by improving lane markings and signals, filling potholes, and building complete streets that separate pedestrians and cyclists from motor vehicles benefits all road users and helps AV operations [14]. Similarly, improving access to broadband may promote equity for all community residents as well as assist in AV implementation. Transitioning to electric vehicles promotes environmental sustainability with community-wide benefits.

Other infrastructure improvements may favor AV users while reducing resources available to improve transportation access for disadvantaged populations. For example, private or public investments in infrastructure-to-vehicle communications technologies (such as traffic lights that emit digital signals), dedicated AV electric charging stations or detailed road mapping capabilities could facilitate expanded use of AVs. Dedicating road lanes for the exclusive AV use would represent an inequitable use of public resources.

Decisions about proposed infrastructure projects designed to benefit AV use should incorporate community input and equity considerations. Some infrastructure improvements to support AVs might be funded by private rather than public resources, like the electric charging stations supported by Tesla.

Equitable provision of AV services

US transportation services have historically discriminated based on income, race and ethnicity [6, 13]. For instance, Black individuals face longer wait times and higher trip cancelations for taxis in Los Angeles and for ride-hail services in Seattle and Boston compared to other racial and ethnic groups [9, 35].

Equity concerns over AV accessibility span income, race, ethnicity, age and regions. In Seattle, an activity-based modeling exercise suggested AVs might enhance perceived accessibility more in rural than urban areas, a trend supported by other studies [27, 36]. Some studies anticipate initial benefits favoring moderate to high-income individuals, potentially impacting social equity and public health [2, 23]. Such studies suggest AVs may provide fewer services for disadvantaged populations, causing or exacerbating adverse impacts on social equity. Policies favoring shared over private AV ownership may improve accessibility [18].

The impact of AVs on racial discrimination is less understood. AV operations rely on large datasets, raising concerns about artificial intelligence (AI) bias if data collection favors high-income or White neighborhoods. Discrimination in AI-driven systems could surpass existing transit system practices, necessitating policies ensuring equity in AV co-ownership and fleet models and detecting discrimination [65, 66].

Other discriminatory practices include concentrating AV testing in low-income communities, which poses risks to disadvantaged groups [19]. Transparent and accountable deployment decisions, with advance notice for local residents, are crucial. Regulations should ensure ethical considerations are uniformly applied, fostering competition among companies committed to ethics [67]. Uncertainty in liability rules, recovery, damages and insurance policies for AV crashes disproportionately burdens lower-income groups and should be addressed [19].

Access to AVs by persons with low incomes and/or lacking smartphones or credit cards

Equitable access to AVs hinges on affordability, particularly in urban settings where low-income individuals often rely on public transit and non-motorized trips [20, 37, 39]. The quality and availability of public transit vary, and funding shifts to AV systems may limit access for low-income individuals, impacting their ability to reach essential services [28, 68, 69]. AV benefits may favor those who can afford ownership or travel costs [38]. Early AV use policies should incentivize access for low-income and disadvantaged populations, especially those with non-traditional work hours that could benefit from 24-hour AV services. Policies favoring large, shared fleets and targeted fare subsidies promote equity and accessibility over private ownership-focused policies.

Shared AVs, akin to current ride-hailing services, will likely be requested by smartphone. In 2021, 85% of US households owned smartphones, more prevalent in urban than rural areas [70]. Those lacking smartphones or credit cards face AV access barriers. Addressing this could involve subsidizing smartphones for low-income households and enabling cash or voucher payments for AV rides. Seniors might benefit from simplified smartphone models with essential apps. Exploring alternative ride-request methods, like AV calling stations at transit stops and commercial areas, could be valuable, although less practical in low-density residential zones. Subsidizing AV use may increase externalities and health impacts associated with heightened AV use, and such funds might better be used in other ways to improve accessibility and reduce externalities for low-income households. However, political realities might prevent funds that could be used for AV subsidies from being spent on alternative uses.

Shared infrastructure services

Existing vehicles and AVs are likely to share most infrastructure, such as roads, traffic signals and signage. Some new infrastructure would be needed to optimize AV implementation that may compete with funding for infrastructure for conventional vehicles, public transit, pedestrians and bicycling.

Road traffic is likely to include both automated and conventional vehicles in the coming decades. Studies of real-world and computer-simulated mixed traffic flows show that efficiency and safety of road transportation could decrease if no measures are taken to support the coexistence of conventional and automated vehicles [71]. AVs are expected to be implemented in phases. AVs might share lanes with buses and high occupancy vehicles. Shared AVs may replace current ride hail services that have drivers. These projections have potential for reducing traffic congestion and air pollution. A study in Austin, Texas using an agent-based model suggests each shared AV could replace around 10 privately owned or household-owned vehicles [40], leading to fewer cars on the road or occupying parking spaces.

Safety concerns as a barrier to shared AV use

Although a person with sufficient resources may prefer private AV ownership [42], more societal benefits will be realized if AVs are owned in fleets and if most AV rides are shared with other persons going to similar destinations. Such shared rides provide cost savings per trip and help reduce vehicle miles traveled, traffic congestion and vehicular emissions. Barriers to shared AV use arise due to concerns over safety especially among women, racial and ethnic minority groups, children, persons with disabilities and LGBTQ+ persons. Safety considerations for all potential AV users need to be addressed in policies for AV implementation.

Safety concerns for women

Personal safety is a primary concern for women that influences their travel behavior. Women do numerous trips such as commuting to work, transporting children and running errands. Land use (car-dependent suburban sprawl vs compact urban design), trip purpose (work commute vs. transporting children or groceries) and time constraints influence whether a trip is practical by transit, walking, or micromobility or needs an automobile. Women have faced discriminatory and uncomfortable treatment from ride hail drivers at times such as more expensive routes and unwanted flirtation [9]. Uber received almost 6000 complaints of sexual harassment (from unwanted touching to rape) by their drivers in 2018 and 2019 [43].

AVs without drivers could eliminate such uncomfortable rider experiences, especially for women. However, 81% women and 67% of men are skeptical of the safety of fully automated vehicles [7]. Women may be afraid to use public transit after dark [7], including time waiting for transit transfers. Such concerns may lead women to consider carpooling only with known persons. Women-only ride hailing options with female drivers are offered by some companies like Sakha Cabs in India, Fyonka’s in Egypt, Lily Ride and Ojek Syari in Indonesia and Bangladesh, Sheebah in Australia, Green Cab in South Africa, Rosy and Pink Cabs in the UK. In Saudi Arabia and Brazil, Uber offers a ‘Women Rider Preference’ feature. [46]. Women-only rideshare services are accessible to trans women in at least one country (Sheebah in Australia).

AVs have the potential to assist caregivers by reducing their driving responsibilities [7] such as transporting teenagers to sports practices or seniors to medical appointments. However, when used in services like UberPool and LyftShare where passengers share rides with strangers, AV shared rides may entail risks in the absence of human drivers who can maintain order in the vehicle.

Safety concerns for children

Guidelines for safe transportation of child passengers in AVs in the absence of adult supervision are needed to address scenarios such as children unbuckling themselves, playing unsafely or traveling alone in AVs. Best practices should be developed considering factors like meeting the child at their destination, handling breakdowns or re-routes, ensuring communication within the vehicle, preventing premature exits and managing co-passengers picked up during the journey [47]. Children-only rideshare fleet services could provide some safety from unwelcome adult passengers but would need to address behavioral supervision.

Safety concerns for people of color

Less sharing of ride-hail services occurs in racial and ethnically diverse neighborhoods than in more homogeneous neighborhoods [44]. Users living in low-income neighborhoods tend to share ride-hail trips for a higher proportion of their trips than those in high-income neighborhoods. Riders with discriminatory attitudes tend to not want to use rideshare again and are less satisfied with rideshare [8]. This suggests a reluctance of people in diverse neighborhoods to share rides with ethnically different persons and may be associated with uncomfortable and possibly dangerous discriminatory behavior.

Addressing safety concerns

AV acceptability among users may be enhanced by onboard safety such as live intercom, and auditory and visual cues for those with hearing and visual impairments [41]. Existing sensor technologies including interior and exterior cameras could be designed to enhance rider safety [45]. Other measures, such as more careful matching of users, remote monitoring through cameras, safety partitions and emergency call buttons could help reduce the vulnerability of riders to inappropriate behaviors from other passengers [7]. The use of interior cameras and recordings can cause privacy concerns and would require user consent and assurance on recording disposition. AVs could provide safe exit strategies for passengers encountering uncomfortable situations. For example, an AV could drop-off a passenger near rather than in front of their home to protect their privacy from a co-passenger [45]. Because voice interaction is most effective in time-sensitive situations, AVs could use verbal commands to trigger safety protocols [45]. AI could be used to detect aggression or harassment behavior with interior cameras and then ban offending riders from future AV services. Another strategy, used by Ola in India (www.olacabs.com), is an option for customers to automatically notify their emergency contacts of ride information whenever they hail the rideshare service.

Access to AVs by persons with disabilities and senior citizens

People with disabilities often struggle to access convenient transit systems [7, 53]. The US transportation system is predominantly dependent on automobiles. Losing the ability to drive means losing transportation independence, affecting those with physical and/or mental disabilities and older individuals [24]. Beyond depending on family, friends or paid assistants, public transit and ride-hail services become some of the only accessible methods of transportation. Challenges with using public transit include accessing transit stations, hearing announcements, reading signs and accessing emergency care during trips. AVs may offer increased independence by reducing wait times, providing first and last mile connectivity and minimizing transfers. Barraza et al. [54] discuss how the Americans with Disabilities Act applies to AV services.

Seniors may be reluctant to use AVs if they are concerned about AV reliability [7, 50]. An interview survey of 63 older adults who interacted with and rode a shared autonomous bus shuttle revealed generally positive reactions. In addition to advocating for accessible ergonomic standards, participants recommended convenient routes and having a human assistant to increase receptivity [72].

Standard AVs may or may not help persons with disabilities, especially those in wheelchairs or who need assistance to enter and exit a vehicle. In addition to paratransit services, policies could mandate in any AV serviced jurisdiction that a percentage of vehicles accommodate wheelchair users and provide aids for hearing and visually impaired users [73]. Policies that mandate targeted subsidies for AV access and technical assistance may promote active aging among older populations by enhancing mobility options. These policies must be informed by studies on wellbeing for aging societies, including an understanding of acceptance factors and the learning curve for use of AV technology among seniors [24, 48, 51].

Disruption of existing jobs in transportation

The advent of AVs may eliminate 1.3 to 2.3 million jobs over the next three decades [27, 55, 56]. While AVs could eventually replace transit drivers, the current driver shortage in many agencies might be eased by their deployment [74]. Jobs linked to widespread AV use will demand different skills and offer varied pay scales compared to current transport-related roles [55, 57]. Occupations like delivery truck, heavy truck, bus, and taxi driving may see substantial declines, affecting a significant source of jobs for those with lower education levels; 93% in these positions have less than a bachelor’s degree [56]. Men and people of color are disproportionately represented in these jobs [56], and their transition to other roles may pose challenges for employees and employers. Widespread AV use could decrease jobs related to driver licensing, insurance sales, traffic policing and parking attendants, and require a modified skillset for mechanic-related jobs [24, 56]. Addressing job losses could involve retraining workers for roles in high demand, like mechanics for electric vehicles and coordinators for AV fleets, although the balance of job loss and gain remains uncertain.

DISCUSSION

A fair transport policy reduces inequities in access to opportunities across social groups through equitable distribution of investments and services [23]. Such policies establish minimum accessibility standards, prioritize disadvantaged groups and address externalities like air pollution and traffic crashes [23]. Inclusive processes involving current/future users, community leaders, stakeholders and disadvantaged communities should frame transport policies [18]. AV policy development requires input from professionals in public health, urban planning, social justice, cybersecurity, AI, transportation planning and AV system operators. Relevant stakeholders include youth, seniors, non-car users, rural residents, low-income individuals and persons with disabilities [33]. Some states are considering equity issues in their planning for future AV use [75].

Communities vary in their demographics, geography and structural inequities, and AV policies need to address different opportunities and challenges based on the context. AV transportation policies will need periodic updates to accommodate changes in technology and infrastructure [24].

Emory et al. [37] published a literature review on equity implications of AV policies across various jurisdictions. They found common policies concerning AV shared use, economic impacts, marginalized communities, urban parking and transit automation. However, few policies addressed issues like low-income and minority groups’ needs, personal security in shared vehicles or rural AV deployment models.

Based on the nine equity issues identified in our review, policymakers should consider developing (i) policies governing how the city or state jurisdiction oversees AV implementation and (ii) policies specific to the permitting process to allow individual companies to offer AV services in their jurisdiction (Table 2).

Table 2

Policy recommendations to improve equity associated with widespread implementation of AVs

Policies governing how a jurisdiction oversees AV implementation

1. Include input from representatives of disadvantaged populations in policy development

2. Ensure affordability of AV use for low-income persons and persons without smart phones or credit cards

3. Ensure equity in selection of and funding for infrastructure improvements

4. Develop guidelines for AV ridesharing that protect vulnerable populations

5. Provide job retraining for taxi, ride hail and truck drivers who may lose jobs from widespread AV implementation

6. Conduct ongoing monitoring and evaluation of AV services from an equity perspective

Policies specific to the permitting process that allows companies to offer AV services in the jurisdiction

1. Review AV service provider plans to accommodate accessibility for persons with mobility, vision and hearing disabilities

2. Review AV service provider safety and security procedures for vulnerable users

3. Assure the privacy of AV users including data collection and retention

4. Review service area proposed by the company from an equity perspective

5. Assess the AI algorithms used to guide the AV services from an equity perspective

Policies governing how a jurisdiction oversees AV implementation

Incorporating equity in AV deployment requires involving representatives of disadvantaged populations in policy and planning. Active outreach is needed to identify persons of color, low-income individuals, and those with disabilities, who may lack knowledge about AVs and need education before providing informed opinions. Education may cover AV operations and scenarios impacting daily activities and may be developed collaboratively by government agencies and private stakeholders. Materials and services must be in multiple languages for non-English speakers. The public input stage should allow communities to highlight their transportation priorities. Some communities may prioritize improved public transit and sidewalks over AV infrastructure improvements.

A second issue is ensuring affordability for persons with low incomes. Approaches include subsidies for the cost of AV rides and provision of alternative means of access to AVs for persons who lack smartphones and credit cards [2, 76]. AV call boxes at commonly used sites, such as transit stops, would allow persons to call for AV services without a smartphone. Some transit systems currently offer discounted rides for low-income persons and seniors, and these might be combined with ‘last-mile’ AV services from transit stops [77]. Shared AV ownership could be encouraged by financial incentives because AVs operated as part of a shared fleet are more equitable and cost-efficient than privately-owned AVs [25, 51, 78].

A third concern is ensuring equity in funding infrastructure improvements. Plans should be reviewed for equity before funding approval to ensure equitable distribution, prioritizing improvements benefiting all road users, like lane markings and pothole patching. Lower priority should be placed on improvements primarily aiding AVs, such as road-to-vehicle communication technologies. Development of high-definition maps needed for efficient AV use should not be limited to high income areas which private AV service providers might find most profitable [79]. Improvements that primarily help AVs might be funded at least in part by the private sector.

A fourth issue is the development of policies to guide ridesharing in AVs. Ridesharing should be encouraged because it contributes to affordability and efficiency but also raises questions about personal safety for vulnerable persons and liability when untoward events occur. Most current trips in ride hail vehicles do not involve ride sharing despite the economic and environmental benefits of shared rides. Substantially lower fares for shared rides compared to solo rides help encourage the use of shared rides.

A fifth issue is the provision of job retraining for persons such as taxi, ride hail and truck drivers whose jobs may be eliminated if AVs come into widespread use. Retraining programs could provide the skills for such persons to fill new jobs such as those needed for the maintenance and dispatch of AVs. In designing retraining programs, extra effort may be needed for workers whose skills, interests and abilities do not align well with the new positions available.

A final concern is the ongoing monitoring and evaluation of AV services to ensure equitable delivery. This could involve periodic community-wide travel surveys incorporating questions on AV access. Queries could address disability access, affordability for low-income individuals, comfort for persons of color and women, and meeting transportation needs in disadvantaged communities. Survey results and anonymized ride data should inform updates to AV policies.

Policies specific to the permitting process that allows companies to offer AV services in the jurisdiction

In addition to policies to govern all AV use, state or local governments may choose to establish a permitting process for each company that wants to offer AVs services in their jurisdiction, such as those now required in some states for AV testing on public roads. Such permitting processes can be designed to ensure that equity is incorporated into the AV services offered [80].

One issue will be to review how the company plans to accommodate accessibility for persons with mobility, vision and hearing disabilities. Few AVs are designed to make it easy for a person in a wheelchair to enter and exit the vehicle. Companies offering AV services could be required to provide vision and hearing accommodations in all AVs and wheelchair-friendly designs in a certain proportion of their AVs.

A second issue would be a review of the safety and security procedures for vulnerable users, especially women, minorities and seniors. Periodically reviewing customer complaints can gauge the adequacy of safety measures. Enhancements like emergency call buttons and vehicle surveillance (audio and video) can improve safety. The Institute of Electrical and Electronics Engineers developed an engineering standard approach to integrate human and social values into engineering design. AV companies may be required to adhere to similar ethical guidelines to consider stakeholder values in product design and systems requirements. These standards provide ethically oriented regulatory obligations for autonomous systems [81].

A third issue is assuring the privacy of AV users. Companies must be required to protect the privacy of data such as pickup and destination addresses, and video recordings made within the vehicle. Regulations are needed to govern the retention and destruction of such records.

A fourth issue would be a review of the service area proposed by the company to ensure that AVs are made available to all areas of the community, not just higher income neighborhoods. A periodic review of company records could document the geographic distribution of the services actually provided. Similarly, low-income areas should not be disproportionately burdened as parking areas for AVs between service calls [37].

A fifth issue would be an assessment by AV regulators of the AI algorithms used to guide the AV services. Frameworks that govern AI use should be fair, transparent and accountable [82]. Companies could be required to document that their algorithms are based on demographic and travel data from diverse populations to ensure equity in AV services. Few (if any) methods now exist to document if an algorithm is equitable. Further research in this area is needed.

Limitations

Several limitations should be considered in interpreting the results of this review. The field of AVs is growing rapidly and the equity aspects of studies now in progress may not have been identified. Companies developing AVs may have relevant proprietary information that is not available for review. This review focused on AVs used primarily by passengers on public roads and excluded other uses of AVs, such as package delivery, truck platooning and AV shuttles serving limited domains such as campuses or airports.

The equity impacts of AI algorithms guiding AVs are hard to gauge, especially for populations not in pilot studies. AVs rely on AI algorithms to detect pedestrians and bicyclists based on size, shape and movement, not facial recognition. It seems unlikely that object recognition algorithms would have inequitable difficulty in detecting persons with dark skin tones, especially because many pedestrians of all races wear dark clothes. The possibility of algorithm-driven racial discrimination was suggested in one non-peer reviewed report [83].

Despite these limitations, we believe that the equity issues highlighted in this review can serve as a basis for discussions leading to more equitable policies related to AV use.

FURTHER RESEARCH

While some road infrastructure improvements for AV users such as better lane markings are likely to benefit other transportation modes, further research could help specify the improvements needed to accommodate widespread AV fleets and to identify which improvements help multiple transport modes versus those that primarily help AV users.

Research is needed to explore the best transportation options and modes for increased equitable transportation services to low-income and underserviced neighborhoods. One key concern is whether ensuring AV access in low-income neighborhoods is the best approach to achieving equitable services compared to other options such as improved transit service and sidewalk networks. Related research could explore the impacts on other transportation modes of road improvements designed to assist AV users. Research could also explore the implications of widespread AV use on horizontal equity (persons with similar incomes pay similar rates for a given service) and on vertical equity (persons with higher incomes pay more for a given service).

Various investigators have predicted that widespread use of AVs may lead to increased urban sprawl, vehicles miles traveled and sedentary lifestyles, and decreased injuries along with inequitably distributed externalities. Future research will be needed to assess the accuracy of these predictions. This review could be repeated using more systematic review methods in a few years, by which time more data on the topic may be available.

There may be adverse health effects associated with AV lithium batteries and with exposure to electromagnetic fields. Further research could be done to better document the magnitude of these potential health effects in the general population and to examine whether vulnerable populations are disproportionately affected by these hazards.

CONCLUSIONS

Widespread use of AVs will have both positive and negative effects on societies, ranging from decreased motor vehicle injuries and more efficient transportation options to inequitable investments that harm disadvantaged populations. Public sector policies informed by professionals in public health, social services, transportation and other fields can help maximize the positive and minimize the adverse impacts on equity of increasing AV use in coming years.

DISCLOSURE

To meet the journal’s word limit after the paper was revised in response to reviewer comments, the authors reduced the word count by ~25% with the assistance of GPT 3.5. The authors reviewed and fact checked all edits suggested by GPT.

Sections rewritten with AI: 3.1; 3.2; 3.4; 3.5; 3.6; 3.10; parts of 4.1 and 4.2.

ACKNOWLEDGEMENTS

We thank William Covington, Frank Douma, and Andrew Raitt for their helpful comments on a draft of this manuscript.

STUDY FUNDING

The authors do not have any funding sources to disclose

CONFLICT OF INTEREST

None declared.

AUTHORS' CONTRIBUTIONS

The authors confirm contribution to the paper as follows: study conception and design: A.L.D.; data collection: S.F., C.H.L., A.L.D.; analysis and interpretation of results: S.F., C.H.L., A.L.D.; draft manuscript preparation: S.F., C.H.L., A.L.D. All authors reviewed the results and approved the final version of the manuscript.

DATA AVAILABILITY STATEMENT

No new data were generated for this report. The information described is a synthesis of information existing in the literature that is referenced throughout the report.