The promise of widespread use of fully autonomous vehicles by private users has yet to materialize. Promoting and implementing the rapid adoption of this technology is evidently more complex than initially imagined. Yet driverless taxi services, or “robotaxis,” are already here. Several U.S. cities now offer such taxi services, with a clear prospect of expansion. In Regulating Robotaxis, Bryant Walker Smith and Matthew T. Wansley seize this moment to analyze the pros and cons of this phenomenon while proposing a regulatory blueprint for future expansion and deployment of robotaxis throughout the United States. The paper is a must read for anyone interested in this emerging business model and its implications.
The paper begins by setting the scene and introducing the basic technologies featured in these vehicles, especially their sensors and software, as well as the roles humans still play in their operation. Next, it maps out the existing robotaxi markets, which are currently dominated by firms linked to today’s digital giants: Waymo, a subsidiary of Alphabet (the parent company of Google); Zoox, a subsidiary of Amazon; and Tesla. This enables the authors to do two things: examine the impact of widespread use of autonomous vehicles, and call attention to the implications of the emergence of autonomous taxi fleets as powerful players at technological, economic, and political junctures.
The article discusses the costs of cars and operations, and potential savings. It also notes open questions, such as whether robotaxi users will tolerate shared rides, while explaining how the car’s interior design might make a difference. The authors then provide a comprehensive review of the strategies applied by various states, such as California and Arizona, to regulate robotaxi fleets, focusing on objectives such as safety and cybersecurity. Some regulations include in-depth disclosures, meeting standards, and passing tests. In addition, states have introduced rules to address fears of price gouging. State regulators have also introduced “universal service” requirements to ensure that a broader segment of the population can benefit from these fleets and their services.
The noted regulations govern a new industry that arises out of two distinct and highly regulated contexts. First, there is regulation of Transportation Network Companies (“TNCs”) such as Uber and Lyft and their taxi predecessors. Second, there is emerging and complex regulation of self-driving, or fully autonomous, vehicles. Robotaxi regulation must figure out how to synthesize rules from both contexts in order to advance social objectives, without stifling innovative technologies and business models.
The paper next turns to a normative analysis. It introduces several forms of externalities generated by extensive robotaxi fleets and considers if, when, and how they should be internalized. The paper addresses pollution, wear-and-tear to infrastructure, congestion, and privacy. The authors point out, for each of these externalities, how the growth of robotaxi usage might generate social and other costs. Then it explores which regulatory steps might mitigate these costs.
The analysis is insightful, although it might be strengthened by examining the potential use of robotaxis for delivery services, a development that could exacerbate some externalities while mitigating others, such as bicycle-related safety concerns. The authors, too, could further consider potential privacy-related benefits of robotaxis, not just harms. While the increasing usage of such services might create substantial databases regarding users’ travel patterns and vehicle-captured surroundings, robotaxis remove the human driver and, thus, eliminate a major vector for the collection of behavioral data now gathered by conventional electric and connected cars. For example, a 2024 New York Times report documented extensive data collection by automobile manufacturers relating to driving behavior, including indicators such as “speeding, hard braking, or sharp accelerations.” This information was reportedly shared with data brokers. It even led to adverse effects on insurance premiums. Moreover, so-called “Driver Monitoring Systems” (DMS) are being deployed in a growing number of vehicles; these systems monitor, among others, driver attention and drowsiness by collecting data on posture, eyelid movement, and related physiological cues. Clearly, such practices might prove harmful if the data makes its way to the wrong hands. Such forms of data collection cease when the individual transitions from the driver’s seat to the passenger seat. In this respect, robotaxis may promote certain privacy interests even as they complicate others.
In Part III, the authors consider “rider protection,” focusing on three broad dimensions: consumer protection, safety, and autonomy. The authors fear that insufficient competition might lead to problems in each of the noted dimensions and discuss regulatory measures to enhance competition by enabling market entry and facilitating switching between services. They are also correct to note that autonomy concerns will emerge in times of emergency, when users might seek to rely on their private vehicle to escape a looming risk, such as a natural disaster. The availability of robotaxi services is unlikely to induce a complete abandonment of private (even gas-operated) vehicle ownership, particularly given the perceived need for immediate access to a reliable functioning vehicle in emergency situations such as floods. However, because many households maintain multiple vehicles, the presence of reliable robotaxi alternatives may encourage a partial divestment of private cars, with households retaining a single vehicle for contingencies. From a policy and sustainability perspective, such a shift would represent a remarkable change in the transportation market.
In its final segment, the article examines broader implications of robotaxis. These include the possible reshaping of urban spaces given lower parking demand, and expanding mobility for low-income riders and people with disabilities, among other effects.
The article’s driving (yes, driving) force is a sense of urgency: the authors’ sense that the issues described above must be dealt with now, because over time the political economy might render this objective unrealistic. As recent decades have shown, emerging tech firms often evolve through growth or mergers into giants with substantial political power. They become too big to regulate or fail. The fact that the budding robotaxi leaders are linked to digital powerhouses contributes to this sense of urgency – and in fact, given the vast power of these entities, we might even be too late to block their influence in the robotaxi sphere. It is also instructive to consider who exited the market and who might soon enter. Cruise, a subsidiary of General Motors and part of the “old guard,” has ceased its robotaxi operations, indicating yet another victory for the new firms over the previous generation. On the other hand, Chinese firms might enter the U.S. market, such as Apollo Go, a subsidiary of Chinese tech giant Baidu. These latter firms might bring with them capital, expertise, and influence with local regulators. We may stand at a fleeting moment for these taxi fleets. Regulators should take account of these issues and this article, without delay.
