Self-Driving Cars Could Worsen Urban Problems—Unless Cities Act Wisely

Self-driving cars, also known as autonomous vehicles (AVs) or robotaxis, promise safer roads, reduced parking needs, and more productive travel time. Companies like Waymo, Cruise, and Tesla are advancing the technology rapidly, with deployments already underway in several cities. Yet, without thoughtful regulation and urban planning, widespread adoption risks amplifying the very issues that already strain modern cities: congestion, sprawl, declining public transit, and streets designed primarily for vehicles rather than people.

This concern echoes analyses from urban planning experts, including the popular video essay by Jason Slaughter of Not Just Bikes. The core problem is not the technology itself but how it interacts with existing incentives and infrastructure. AVs could function as an efficient supplement to multimodal transport—or they could entrench car dominance on a larger scale.

Induced Demand and Rising Vehicle Miles Traveled

One of the biggest risks is induced demand. When travel becomes cheaper, easier, and more convenient—thanks to no driver costs, the ability to work or relax en route, and potentially lower per-mile expenses—people tend to travel more often and farther. This is the same dynamic observed with highway expansions: added capacity quickly fills with new trips.

Meta-analyses and simulations of AV deployment in U.S. cities show vehicle miles traveled (VMT) increasing by roughly 6% or more overall. Non-shared private AVs tend to generate higher increases (around 7%) compared to shared fleets, though both contribute to growth. Factors driving this include reduced perceived travel time costs, new users (such as the elderly or those previously reliant on transit), longer commutes, and modal shifts away from buses and trains.

Early data from partial automation features, like Tesla Autopilot, already suggest owners drive thousands of additional miles per year. Full autonomy could magnify this effect, overwhelming efficiency gains from smoother traffic flow and leading to worse congestion in many scenarios.

Empty Miles and Deadheading

Robotaxis introduce another inefficiency: deadheading—trips made without passengers as vehicles reposition, return to depots, or cruise for the next fare. Real-world operations in places like San Francisco show deadhead miles accounting for 40-50% of total travel in some periods. These empty movements consume road space, generate tire and brake wear (contributing to pollution even in electric vehicles), and add noise without moving anyone.

Simulations indicate empty VMT could represent 15-25% or more of fleet operations, depending on fleet size, penetration rate, and management. Without curbs on such trips, cities could see streets increasingly occupied by circulating vehicles rather than productive ones.

Threats to Public Transit and Active Mobility

Cheap, on-demand robotaxis could accelerate the decline of public transit. Door-to-door service competes directly with buses and trains, especially if subsidized by venture capital or operated at scale to capture market share. As ridership falls, transit agencies face revenue losses, leading to service cuts and a vicious cycle of further decline.

This mirrors how automobiles historically marginalized walking, cycling, and transit in many North American cities. AVs might extend this pattern unless integrated thoughtfully as feeders to high-capacity rail or bus rapid transit rather than full replacements. Vulnerable populations who rely on affordable transit could suffer most if service erodes.

Sprawl, Space Inefficiency, and Street Redesign Pressures

Easier long-distance travel encourages urban sprawl. People may choose more affordable housing farther from job centers, knowing commutes can be productive or restful. Models show mixed outcomes: reclaimed parking land could support denser downtowns, but lower effective commuting costs often push development outward, increasing overall energy use and infrastructure demands.

Even autonomous, cars remain space-inefficient for moving people. A single-occupancy vehicle (or empty one) occupies far more road and curb space per passenger than a full bus, bike lane, or sidewalk network. Cities may face pressure to prioritize AV flow—higher speeds, fewer traffic signals, removal of bike/pedestrian infrastructure, or widened roads—further eroding human-scale urban environments.

Studies highlight tensions between these forces. In some projections, unchecked AV growth leads to more congestion, longer trips, and higher emissions despite electrification. Noise and particulate pollution from tires and brakes persist as concerns.

What Cities Should Do: Policy Over Technology Determinism

The technology is not inevitable destiny. Cities shaped themselves around cars in the 20th century through deliberate policy choices; similar choices today can steer AVs toward better outcomes.

Key recommendations include:

• Treat robotaxis as a regulated public utility. Implement per-mile fees, especially for empty/deadhead trips, congestion pricing, fleet caps in dense areas, and geofencing rules. Require integration with public transit apps and schedules to encourage complementarity rather than competition.
• Prioritize people-first infrastructure. Invest aggressively in safe walking, cycling networks, and high-quality transit before AV dominance takes hold. Promote dense, mixed-use development that keeps most daily trips short and walkable or bikeable.
• Address land use and pricing. Tax or regulate sprawl-inducing development. Rapidly repurpose parking land for housing, parks, or active uses. Ensure the full societal costs of car use—including road space, pollution, and congestion—are reflected in pricing.
• Learn from successful models. European and Dutch-style cities demonstrate that strong protections for non-car modes, combined with land-use policies favoring density, create livable environments where AVs can enhance mobility without overwhelming it. AVs work best as last-mile connectors in such systems.

Evidence from modeling shows that policies like road pricing, shared-use mandates, and zero-emission requirements can significantly mitigate VMT growth and negative externalities. Shared fleets with high occupancy, paired with supportive urban design, yield better results than private AV ownership.

In summary, self-driving cars offer real benefits—improved safety for vulnerable road users, better mobility for those unable to drive, and potential reductions in parking footprints. Realizing these gains while avoiding amplified car dependence requires proactive governance. Cities that regulate early, protect multimodal options, and design for humans rather than vehicles will be better positioned in an autonomous future. Those that treat AVs as an unregulated panacea risk locking in less livable, more congested, and more expensive urban environments.

The coming years represent a critical window. Technology will advance regardless, but its urban impact depends on the policies we choose now.

About The Author