By Jonathan Smith and Prof Siddartha Khastgir
A PAVE United Kingdom publication
Introduction
Engineering excellence alone cannot guarantee the success of autonomous vehicles (AVs). While technical hurdles like "edge cases" dominate industry discourse, the ultimate adoption of self-driving technology hinges on building a "trustworthy system" through deep, meaningful public engagement. This paper synthesizes over 100 studies to provide a strategic playbook for moving from one-way technical communication to collaborative co-creation, ensuring technology aligns with public values.
The Problem: Addressing the Trust Gap
History demonstrates that technically impressive innovations often fail when they ignore public sentiment. A recent example is the halting of "Smart Motorways" in the UK. The core challenge lies in a mismatch between expert assumptions and public perception. When designers misjudge public understanding, it leads to three critical failures:
- Misuse: Over-reliance on a system's capabilities beyond its design.
- Disuse: Skepticism and rejection of the technology.
- Abuse: Intentional interference with the system by the public.
To bridge this "trust gap," the industry must move beyond marketing hype toward an evidence-based framework for public acceptance before these vehicles are deployed at scale.
Methodology
This narrative review synthesizes evidence across multiple domains to build a comprehensive understanding of trust:
- Case Study Analysis: Examining events like the DeepMind/NHS data breach and UK nuclear policy to understand how public disagreement ("sites of contention") can actually be used to strengthen governance.
- Theoretical Modeling: Adapting trust models (such as Lee and See, 2004) to distinguish between Trust in the system (pre-use beliefs) and Trust with the system (real-time interaction).
- Meta-Analysis: Reviewing over 100 studies to identify psychological drivers of sentiment, utilizing the "63 Factors" Framework to categorize user and service-related variables.
Key Insights: Trust as a System-of-Systems
Trust is not merely about sensor accuracy; it is a triad of Technology, People, and Policy. A failure in transparent data handling or policy can collapse the entire system. Furthermore, the "Information Deficit" fallacy suggests that simply providing more technical data does not increase trust. Instead, engagement must be symmetrical—a two-way street where public feedback directly influences design.
The goal is not simply "high trust," but calibrated trust: ensuring users trust the system exactly as much as its current capabilities warrant—no more, no less.
Drivers of Public Acceptance
The review identifies five foundational factors that consistently influence public behavior and adoption:
- Performance Expectancy: The perceived benefit to the user (the strongest predictor of behavior).
- Effort Expectancy: Ease of use. For instance, the presence of a human conductor on an autonomous bus significantly alters this perception.
- Hedonic Motivation: The emotional experience (fun vs. comfort), particularly for younger demographics.
- Attitude: The balance of favorable evaluations, often tied to the perceived morality of collision algorithms.
- Perceived Trust: The overarching attribute that determines final adoption.
Future Work and Limitations
Despite the breadth of current research, the authors identify a "Realism Gap." Most data currently comes from text-heavy surveys or controlled, slow-moving trials rather than high-speed urban environments. Future research must address:
- Vulnerable Road Users (VRUs): How pedestrians and cyclists interact with AVs without traditional human gestures (e.g., a nod or wave).
- Longitudinal Dynamics: Tracking how trust evolves over time as users move from "intention to use" to actual daily usage, including how trust recovers after system malfunctions.
Summary updated