When Software Meets the Road: Responsibility for Defective Smart Cars in the MVP Era
56 Pages Posted: 21 Jul 2023
Date Written: July 14, 2023
This Article reveals and analyzes the rising dominance of the Minimal Viable Product (MVP) dynamic in the car industry and its legal and policy implications, especially given the growing automation trends in the driving experience. The MVP concept refers to releasing products as soon as possible and gathering feedback from early adopters to improve the product on the fly, based on users’ inputs. Many software-related products we use today were introduced at such an unripe stage and were developed into their current versions, given business trends originating from Silicon Valley. As the use of software in cars has grown, the MVP concept has migrated into the vehicle industry.
The experimental nature of software development may require a different legal approach when dealing with vehicles, given the substantial negative impact of traffic accidents. While banning unripe products from the roads seems to be a trivial reaction, different reasons deem this to be an unpractical response. The domination of AI-based technologies prevents a thorough examination of user-computer responses until they hit the road. The shortage of current legal remedies to adequately solve the challenges of the MVP concept requires an immediate rethinking of the regulatory environment to address vehicles with various forms of automation while accounting for both deontological and utilitarian challenges.
The Article starts by describing the software industry’s “Agile” methods and introduces the MVP concept, while providing examples of its success and failure. Next, it demonstrates and provides examples as to how this concept migrated into the car industry. Thereafter, it turns to address the deontological challenges involved with applying the MVP method and argues that not only does the use of technologies in such an early stage require meaningful consent, but also approval of an ethics board in an Institutional Review Board (IRB)-like process.
Next, the Article focuses on a utilitarian analysis, after first establishing that the current measures to limit risky vehicles on the road are insufficient. To do so it addresses reliance on recalls, ex-ante supervision, and regulation of software distribution. It points to the complexities of the driving experience which might only be properly addressed through substantial usage and thus ex post regulation. It further indicates that resolving the noted regulatory tensions by relying on stricter ex post measures premised on the law as it stands today might fail to achieve an optimal outcome. This is due to information asymmetries which might substantially hinder effective litigation by plaintiffs.
The Article concludes by providing some initial proposals to mitigate these latter concerns and enable ex post measures. At the end of the day, because MVP products are likely to stay, the Article concludes that some reforms to liability doctrines might be of the essence. These should include mandating manufacturers to record and report their cars’ behavior and rethinking evidentiary burdens.
Keywords: smart transportation, artificial intelligence, autonomous vehicles, recalls, software liability, human experimentation
Suggested Citation: Suggested Citation