Realising Meaningful Human Control Over Automated Driving Systems: A Multidisciplinary Approach

de Sio, Filippo Santoni; Mecacci, Giulio; Calvert, Simeon; Heikoop, Daniel; Hagenzieker, Marjan; van Arem, Bart · 2022 · Crossref

DOI: 10.1007/s11023-022-09608-8

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This paper presents a multidisciplinary framework for realizing "meaningful human control" (MHC) over Automated Driving Systems (ADS), synthesizing research from philosophy, behavioral psychology, and traffic engineering conducted at Delft University of Technology between 2017 and 2021. The study addresses the ethical and legal problem of "responsibility gaps"—situations where undesirable outcomes occur, but it is unclear who is morally or legally responsible due to the opacity and complexity of automated systems. The authors argue that human persons and institutions, rather than algorithms, must remain ultimately in control of driving operations to ensure safety and accountability, particularly as the industry transitions toward partial autonomy and supervised automation rather than immediate full automation. The core of the proposed framework rests on two philosophical conditions for MHC: "tracking" and "tracing." Tracking requires that the ADS’s behavior aligns with the relevant moral reasons and intentions of human actors. The authors operationalize this through a "proximity scale of reasons," which correlates agents and their intentions with system behavior based on temporal and spatial proximity. Tracing requires that at least one human agent possesses the cognitive, physical, and moral capacity to understand the system and accept responsibility for its actions. This condition is operationalized via an "evaluation cascade table," a tool using a six-point Likert scale to assess four aspects: operational control exertion, human involvement, system understanding, and moral responsibility awareness. The final score determines the degree of traceability, ensuring that control is not merely technical but morally grounded. From a behavioral perspective, the paper analyzes the mismatch between driver capabilities and ADS demands, particularly at SAE Level 3 automation. The authors identify an "unsafe valley of automation" where drivers are required to monitor systems for extended periods—a task humans perform poorly—while remaining the fallback for emergencies. Focus group discussions with driver examiners revealed that current driver education fails to ensure proper understanding of ADS functionalities, exacerbating responsibility gaps. The study highlights that without aligning driver training and human-machine interfaces with these new tasks, drivers may end up in a "moral crumple zone," blamed for outcomes they could not control. The significance of this work lies in its translation of abstract ethical principles into actionable design criteria for engineers and policymakers. By integrating philosophical norms with empirical tools like the evaluation cascade and proximity scale, the framework provides a method to expose deficiencies in traceability and avoid responsibility gaps. The authors conclude that realizing MHC requires not only technical adjustments but also institutional changes, including revised driver education and supervisory control protocols, to ensure that human accountability remains intact in mixed traffic environments.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-nvidia on 2026-08-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• automation
• automation surprise
• driverless ads
• automation complacency bias
• situational awareness
• trust calibration

Information type

What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).

• Theoretical Contribution: conceptual framework, computational model, theory or model