Trials and tribulations: Learning from and governing urban experiments with self-driving vehicles

Keywords: Automation, self-driving cars, science and technology studies, smart cities, technology policy
Abstract Type: Paper Abstract

Abstract

Self-driving (or ‘autonomous’) vehicles (AVs) represent a profound test case for artificial intelligence in open, unpredictable contexts. Many cities have played host to trials of self-driving car technology, offering up their roads as a form of testbed. But what is on trial and what is really being learnt? This chapter reports on our experiences with AV trials. These trials, on their own, are a poor indication of possible urban futures. They are typically designed to give the impression that the technology is unconstrained, which means they are intentionally disconnected from questions of governance. Our conclusion is that, for trials to be made relevant for cities, planners and policymakers must play a more active role in shaping and learning from them.

In addition, we will present some of the recommendations of a recent report authored by Stilgoe and colleagues for the UK Government’s centre for data ethics and innovation (https://www.gov.uk/government/publications/responsible-innovation-in-self-driving-vehicles) that seeks a framework that might govern the roll-out and scale up of AV tests.