Virtual assistants and self-driving cars

Self-driving cars are technologically a reality and in the next decade they are expected to reach the highest level of automation. While there is general agreement that an advanced human-autonomous vehicle (HAV) interaction is key to achieve the benefits of self-driving cars, it is less clear what role artificial intelligence (AI) should play in this context. While the scientific community is debating on the role and intersections of AI, autonomous vehicles and related issues, above all ethics, the automotive industry is already presenting AI-based products and services that may influence, in a direction or in another, our technological and societal futures. This paper focuses on virtual assistants, the personification of the car intelligence incorporating, among others, an algorithmic “brain”, a synthetic human “voice” and powerful sensor-based “senses”. Should virtual assistants just assist humans or replace them whenever necessary? Should their scope of action be limited to safety-related driving tasks or to any activity performed in the car or controlled from the car? Although at a very early stage of commercial development, the paper will review the state-of-the-art of in-car virtual assistants underlining their role and functions in the connected and automated driving ecosystem. By drawing from earlier reflections on automation, robots and intelligent agents, it will then identify a series of issues to be addressed by the scientific community, policy-makers and the automotive industry stakeholders.