On the evening of March 18, 2018, Rafaela Vasquez had just started her shift as a ‘vehicle operator’ in an Uber autonomous vehicle. The vehicle was driving test loops on public roads in Tempe, an inner suburb of the city of Phoenix, Arizona. Elaine Herzberg was walking her bicycle across a road when she was struck by the Uber vehicle and died of her injuries later that evening. Within ten days of the collision, Uber completed a financial settlement with the Herzberg family. Two months later, Uber’s autonomous vehicle program left the jurisdiction of Arizona. Uber continued this program in Pittsburgh and San Francisco until selling its autonomous vehicle research unit to Aurora Innovations in December 2020.

     In August 2020, the Maricopa County Attorney’s Office (MCAO), which has prosecutorial jurisdiction over criminal events occurring in Phoenix, charged Vasquez with negligent homicide for the death of Elaine Herzberg. An earlier review, in March 2019, by the neighbouring Yavapai County Attorney’s Office (YCAO), cleared Uber of any criminal responsibility for the fatal collision.

     The burst of media and academic interest triggered by this fatal collision has now dulled. The excited commentary about where liability falls when a person and machine share driving responsibilities has fatigued. More recently, the true consequence of accountability for this collision has been realised, with the criminal proceedings against Vasquez concluding with the parties agreeing to settle the matter. On July 28, 2023, Vasquez pleaded guilty to the reduced charge of ‘endangerment’ and was sentenced to three years’ probation.

     As the only party to be held accountable for this collision, the conclusion of these criminal proceedings against Vasquez now presents a liability crossroads for how society is prepared to attribute criminal responsibility for harms caused by autonomous digital systems and by the corporations who develop this technology. These proceedings can either be accepted as the inevitable result of tragic circumstances or be questioned further to understand better the complexities of who was accountable for what happened. Taking the latter course, not only are there lessons to be learnt about applying laws to autonomous digital systems; there are also lessons about the power of Big Tech and other large corporations and the influence such corporations can have over State agencies charged with regulating these companies.

     In this paper, I will explore the influence Uber had over the Tempe Police Department, as police investigated the fatal collision, and over the offices of the Maricopa and Yavapai County Attorneys as prosecutorial decisions were made, as a case study in regulatory capture. This paper will begin by considering how innovative technologies and corporations developing them impact the traditional concept of regulatory capture. This paper will then introduce the concept of secondary agency capture and explain how the traditional model of agency capture needs to evolve in order to accommodate secondary agency capture. The paper will focus on the particular features of Big Tech that are increasing the risks of capture and the regulatory agencies most likely to be vulnerable to it. The Tempe Police investigation into the death of Elaine Herzberg will then be used as a case study to demonstrate this.

     The paper will also explore the findings of the National Transportation Safety Board (NTSB) to demonstrate the narrow focus taken by Tempe Police and other state agencies toward accountability for the crash. The paper will conclude with a discussion and recommendations.