The Sense of Agency in Driving Automation

Wen Wen, Yoshihiro Kuroki, Hajime AsamaFrontiers in Psychology 2019 · DOI

Abstract

Driving automation has been developing rapidly during the latest decade. However, all current technologies of driving automation still require human drivers' monitoring and intervention. This means that during driving automation, the control by human driver and by the driving automation system are blended. In this case, if the human driver loses the sense of agency over the vehicle, he/she may not be able to actively engage in driving, and may excessively rely on the driving automation system. This review focuses on the subjective feeling of agency of the human driver over the vehicle in such situations. We address the possible measures of agency in driving automation, and discuss the insights from literatures on the sense of agency in joint control, robotics, automation, and driving assistance. We suggest that maintaining the sense of agency for human driver is important for ethical and safety reasons. We further propose a number of avenues for further research, which may help to better design an optimized driving automation considering human sense of agency.

Introduction: The Challenge of Blended Control

We are currently in the era of "blended control" (SAE Levels 1–3), where vehicles handle specific operational tasks while humans remain legally and practically responsible for the driving outcome. However, a well-known "automation irony" persists: as automation becomes more reliable, human operators are increasingly likely to disengage cognitively.

This disengagement is rooted in the psychological mechanism of the Sense of Agency (SoA)—the subjective feeling that one's own actions are causing external events. When a driver no longer feels that their steering or braking inputs directly cause the vehicle's behavior, they lose this sense of agency. This loss leads to over-reliance on the system and dangerously delayed reaction times during critical transitions of control. The central challenge for future automotive design is measuring this invisible mental state in real-time and creating automation that assists without eroding the driver's feeling of being in charge.

Quantifying the Invisible: Measuring Agency

To address this, researchers synthesize findings across cognitive science and engineering to evaluate three primary measurement strategies:

  1. Explicit Measures: Self-reporting through questionnaires. While straightforward, these are prone to bias and interrupt the flow of the driving task.
  2. Implicit Measures: These include Intentional Binding—where the perceived time between an action and its effect feels shorter when one feels in control—and Sensory Attenuation, the brain’s tendency to mute the intensity of self-produced sensations.
  3. Physiological Signals: Electroencephalogram (EEG) data provides objective markers. Specific brain signatures, such as the attenuation of N1 and P3 event-related potentials (ERPs), correlate strongly with a high sense of agency.

Empirical Evidence: Brain Signatures of Control

Research into EEG signatures has shown that when individuals feel in control, the brain's response to an event (the N1 potential) is significantly smaller. This occurs because the brain has already "predicted" the outcome of its own action, leading to sensory attenuation.

Figure 1: ERP waveforms (A) and N1 amplitude (B) across different conditions, illustrating sensory attenuation during self-generated actions.

As illustrated in Figure 1, the N1 amplitude is attenuated during predictable actions compared to unpredictable ones or initial tones. This physiological marker serves as a vital proxy for understanding how automation levels impact the driver's internal model of control. For instance, studies on cruise control demonstrate that drivers exhibit significantly longer braking reaction times compared to manual driving, likely due to the lack of motor preparation that accompanies a high sense of agency.

Conceptual Insights: The "We-Mode"

A surprising insight from recent literature is that SoA is not solely dependent on low-level motor feedback. If a vehicle performs an action that aligns with a driver's high-level goal (e.g., swerving to avoid an obstacle), the sense of agency can remain high even if the specific movement was not initiated by the human.

This suggests that successful automation should foster a "we-mode"—a state of joint action where the human and the AI share intentions. If the AI is perceived as a collaborative "co-pilot" rather than an opaque "black box," the driver is more likely to remain engaged and ready to intervene.

Future Directions in Automotive Design

Moving forward, the focus must shift from laboratory settings to the continuous, high-stakes environment of real-world driving. Key areas for development include:

  • Real-time Monitoring: Utilizing onboard sensors (such as gaze-tracking cameras and steering wheel force sensors) to estimate agency without bulky EEG equipment.
  • Probing for Agency: Implementing "probe stimuli"—minor, safe system deviations—to test if the driver is attentive enough to correct them.
  • Designing for Agency: Engineers should prioritize "trajectory assistance" (helping the driver stay on a chosen path) over "goal assistance" (taking over decision-making), as the former preserves the proximal feedback necessary for a robust sense of agency.

Summary updated

Abstract

Driving automation has been developing rapidly during the latest decade. However, all current technologies of driving automation still require human drivers' monitoring and intervention. This means that during driving automation, the control by human driver and by the driving automation system are blended. In this case, if the human driver loses the sense of agency over the vehicle, he/she may not be able to actively engage in driving, and may excessively rely on the driving automation system. This review focuses on the subjective feeling of agency of the human driver over the vehicle in such situations. We address the possible measures of agency in driving automation, and discuss the insights from literatures on the sense of agency in joint control, robotics, automation, and driving assistance. We suggest that maintaining the sense of agency for human driver is important for ethical and safety reasons. We further propose a number of avenues for further research, which may help to better design an optimized driving automation considering human sense of agency.
BibTeX
@inproceedings{wen2019sense,
  author = {Wen Wen and Yoshihiro Kuroki and Hajime Asama},
  title = {{The Sense of Agency in Driving Automation}},
  booktitle = {Frontiers in Psychology},
  year = {2019},
  doi = {10.3389/fpsyg.2019.02691},
}
Figure