Pilot project purgatory? Assessing automated vehicle pilot projects in U.S. cities

Automated vehicles (AVs) have evolved rapidly over the past decade, prompting cities to explore how to incorporate them into planning. AVs are being tested in many U.S. cities and some are in commercial use. Cities commonly use pilot projects to assess how AVs may help meet transport goals, gauge public interest, and evaluate use cases. Industry narratives promote AVs as transformative for safety, congestion, and mobility, while other scholars question techno-solutionist assumptions and highlight social and political dimensions. Few U.S. cities have incorporated AVs into long-range plans; many either avoid early planning due to uncertainty or develop limited policies or testing rules. Pilot projects are the most common tool. Pilots offer opportunities for real-world testing and potential learning for cities, but risk constraining public discourse by framing AV futures as technical experiments. This paper asks: What are the major trends among cities using AV pilots as a planning tool, and are these approaches likely to positively impact planning? Drawing on interviews and document analysis from 20 cities (from a broader sample of 58, of which 27 used pilots), we examine why cities develop AV pilots, their motivations, and how pilots are anticipated to inform future planning actions.

The paper situates AVs within the broader smart city imaginary where technological progress is linked to sustainability and efficiency, often governed technocratically and critiqued for reproducing inequities and narrowing public debate. Urban experimentation and pilot projects are central tools in smart mobility, yet political biases and normative assumptions embedded in experiments are often overlooked. Much AV research focuses on forecasting implications for travel behavior, land use, parking, and adoption, typically presuming eventual AV deployment and treating AVs as a technical fix. Policy/regulatory literature identifies multi-level governance gaps in the U.S. and offers guidance areas (curb management, transit integration, zoning/parking, data sharing, and pricing). MPOs have begun including AV-related policies, though planners report resource constraints. Cities are using various approaches: regulations, ordinances, mobility playbooks, and pilots. Reviews of AV shuttle pilots show significant variation in location, service, and stakeholder structures; most run in controlled environments and are not fully integrated with transit; participants are not representative of the broader public. Broader critiques of urban experiments highlight “projectification”—short-term, narrowly technical trials with poor linkage to policy learning and limited scalability—often serving industry-led techno-politics and closing down alternative pathways. These insights motivate examination of whether AV pilots are effective tools for planning and policy learning aligned with broader urban goals.

The study combined document/plan analysis and semi-structured interviews. Sampling began with Bloomberg Philanthropies’ AVs in Cities observatory (as of Aug 2019) and was refined to 58 U.S. cities engaged in AV planning, spanning populations from ~10,000 to 8.5 million and involving diverse actors (municipalities, MPOs, transit agencies, universities, state DOTs, private sector). Initial document analysis included city websites, transportation and comprehensive plans, RFPs, media/press releases. Activities grouped into four categories with limited overlap: regulation, testing, planning, and piloting; pilots were most common (27/58). Interviews were conducted in Fall 2019 with 24 participants across 20 cities (IRB-approved; informed consent). Interview guide covered: (1) background (initiators, timing, obstacles); (2) project structure (partnerships, planning context); (3) goals/objectives and alignment with community/transportation goals; and (4) assessment/evaluation. Transcripts underwent content analysis to identify themes. Tables summarizing pilots/plans documented lead organizations, partners, routes/locations, durations, funding, and testing status. The analysis emphasized cross-case thematic insights and illustrative quotations.

• Disconnect between pilot and transportation goals: Interviewees frequently cited city transportation goals (e.g., improving safety: 12 cities; reducing congestion: 13; increasing mobility for disabled/elderly: 5; equity: 5; accessibility: 3; jobs access: 6; compact development/parking reduction: 2). Yet pilots mainly aimed to: introduce AV tech to the public (explicitly cited by 9 cities), test feasibility for transit (often led by transit agencies), identify use cases/infrastructure needs, learn how the tech works, promote economic development, or build institutional capacity. These goals often benefited AV companies more than cities and had weak links to solving transportation challenges.
• Lack of long-term planning: Only 9 of 20 interviewed cities mentioned AVs in long-range transportation or general plans, often superficially. Many pilots were grant-driven, short-term, and ended without continuity. Notable exceptions included Austin, TX, which linked early pilots to a Smart Mobility Roadmap and Strategic Mobility Plan, and Grand Rapids, MI, which co-funded a 12-month pilot with plans for expansion and on-demand services (launched summer 2021).
• Emphasis on non-transportation (economic development) benefits: Several cities pursued AV pilots to brand themselves as innovative hubs or to catalyze economic clusters (e.g., Chandler, AZ; Peachtree Corners, GA’s Curiosity Lab). Claims of economic development benefits often lacked evidence and reflected a pro-innovation bias and “innovation imperative,” potentially crowding out alternative values.
• Limited evaluation and policy learning: About half of completed/ongoing pilots produced a summary report—higher than the 11% assessment rate reported for new mobility pilots generally. However, collected data focused on operational metrics (ridership, reliability, speed, battery, safety events, autonomy percentage) and rider perceptions, offering limited insight into whether AVs advance city transportation goals or inform policy. Some cities aimed to inform state/federal regulation (e.g., San Jose), but most exhibited uncertain links from pilots to policy.
• Insufficient public benefit and data leverage: Despite public investment, data-sharing agreements were often minimal or absent, with benefits (data/learning) accruing primarily to AV vendors. An exception was San Jose, which sought richer sensor-derived insights (e.g., midblock pedestrian crossing heat maps) to guide safety investments. Cities faced challenges aligning infrastructure investments with shifting technology standards (e.g., DSRC vs. 5G in Grand Rapids). Overall, pilots frequently served company R&D more than public value creation.

The findings indicate that current AV pilot practices—short-term, grant-driven, operationally focused experiments—do not effectively serve cities’ stated transportation objectives or long-term planning needs. This addresses the research questions by revealing dominant trends: pilots prioritize public exposure to technology, vendor needs, and economic branding over measurable progress on safety, congestion, equity, or accessibility. The weak integration of AVs into long-range plans and limited policy learning reflect the broader “projectification” of smart mobility, where experiments lack robust evaluation and scaling strategies. The results align with critiques of smart urbanism and techno-solutionism: AV pilots are framed as technical trials, marginalizing political debate about values, equity, and alternative mobility futures, and potentially reinforcing automobility rather than transforming it. Cities that adopt anticipatory governance and proactive policy development (e.g., Seattle, Pittsburgh, Boston) demonstrate a more balanced approach, aiming to shape regulation and align technology with community goals. Strategic partnerships, richer data-sharing, and explicit linkage of pilot metrics to transportation outcomes can shift pilots toward public value. Nevertheless, without embedding AV considerations in long-range planning and articulating clear public-benefit frameworks, pilots risk entrenching industry-led imaginaries and foreclosing more equitable and sustainable mobility pathways.

Current AV pilot practices in U.S. cities—characterized by weak alignment with transportation goals, limited long-term vision, emphasis on non-transportation benefits, insufficient evaluation, and under-leveraged public benefits—provide limited value for informing transport policy and planning. These approaches can reinforce automobility and smart city techno-solutionism rather than enabling transformative change. Cities should: (1) incorporate AVs into long-range transportation plans (via stand-alone strategies or major plan updates) grounded in community goals; (2) design pilots with clear, measurable transportation outcomes tied to safety, congestion, equity, accessibility, and multimodality; (3) establish robust evaluation frameworks and iterate based on findings; (4) negotiate partnerships and data-sharing that yield actionable public benefits (e.g., safety analytics, curb management insights) and protect privacy; and (5) adopt more deliberative, anticipatory, and participatory governance to shape AV deployment. Recognizing that the AV industry needs cities more than the reverse, municipalities have substantial leverage to steer automation toward socially beneficial outcomes rather than following industry-led trajectories.