The attachments of 'autonomous' vehicles

Area of Science:

• Human-Computer Interaction
• Sociotechnical Systems
• Artificial Intelligence Ethics

Background:

• The concept of self-driving cars (AVs) often portrays them as infallible, AI-driven replacements for human drivers, promising liberation from automobility's drawbacks.
• This narrative overlooks the complex web of social and technological dependencies these vehicles will have on their environment and users.

Purpose of the Study:

• To investigate the multifaceted attachments of autonomous vehicles (AVs) to the world, moving beyond the idealized narrative of complete autonomy.
• To explore how AV developers and stakeholders perceive and engage with these attachments.

Main Methods:

• Qualitative study involving 50 in-depth interviews with AV developers, researchers, and other industry stakeholders.
• Analysis of stakeholder perspectives on the social and technological dependencies of AVs, encompassing in-vehicle systems, road infrastructure, and broader societal factors.

Main Results:

• Stakeholders identify diverse attachments, including software, hardware, other road users' behaviors, and the supporting infrastructure (material, social, economic).
• Innovators' approaches to these attachments are categorized into three types: 'brute force' (solving with data), 'solve the world one place at a time' (limiting reach), and 'reduce the complexity of the space' (using attachments as solutions).

Conclusions:

• Autonomous vehicles (AVs) are deeply intertwined with existing social and technological systems, challenging the notion of pure autonomy.
• Understanding these attachments is crucial for anticipating the diverse forms and functionalities future AVs may adopt and for navigating their integration into society.