The Sense of Agency in Driving Automation

Prime-induced illusion of control: The influence of unconscious priming on self-initiated actions and the role of regression to the mean

To what degree human cognition is influenced by subliminal stimuli is a controversial empirical question. One striking example was reported by Linser and Goschke (2007): participants overestimated how much control they had over objectively uncontrollable stimuli when masked congruent primes were presented immediately before the action. Critically, however, unawareness of the masked primes was established by post hoc data selection. In our preregistered study we sought to explore these findings while adjusting prime visibility based on individual thresholds, so that each participant underwent both visible and non-visible conditions. In experiment 1, N = 39 participants engaged in a control judgement task: following the presentation of a semantic prime, they freely selected between two keys, which triggered the appearance of a colored circle. The color of the circles, however, was independent of the key-press. Subsequently, participants assessed their perceived control over the circle's color, based on their key-presses, via a rating scale that ranged from 0 % (no control) to 100 % (complete control). Contrary to Linser and Goschke (2007)'s findings, this experiment demonstrated that predictive information influenced the experience of agency only when primes were consciously processed. In experiment 2, utilizing symbolic (arrow) primes, N = 35 participants had to rate their feeling of control over the effect-stimulus' identity during a two-choice identification paradigm (i.e., they were instructed to press a key corresponding to a target stimulus; with a contingency between target and effect stimulus of 75 %/25 %). The results revealed no significant influence of subliminal priming on agency perceptions. In summary, this study implies that unconscious stimuli may not exert a substantial influence on the conscious experience of agency, underscoring the need for careful consideration of methodological aspects and experimental design's impact on observed phenomena.

Emergence and collapse of reciprocity in semiautomatic driving coordination experiments with humans

Forms of both simple and complex machine intelligence are increasingly acting within human groups in order to affect collective outcomes. Considering the nature of collective action problems, however, such involvement could paradoxically and unintentionally suppress existing beneficial social norms in humans, such as those involving cooperation. Here, we test theoretical predictions about such an effect using a unique cyber-physical lab experiment where online participants (N = 300 in 150 dyads) drive robotic vehicles remotely in a coordination game. We show that autobraking assistance increases human altruism, such as giving way to others, and that communication helps people to make mutual concessions. On the other hand, autosteering assistance completely inhibits the emergence of reciprocity between people in favor of self-interest maximization. The negative social repercussions persist even after the assistance system is deactivated. Furthermore, adding communication capabilities does not relieve this inhibition of reciprocity because people rarely communicate in the presence of autosteering assistance. Our findings suggest that active safety assistance (a form of simple AI support) can alter the dynamics of social coordination between people, including by affecting the trade-off between individual safety and social reciprocity. The difference between autobraking and autosteering assistance appears to relate to whether the assistive technology supports or replaces human agency in social coordination dilemmas. Humans have developed norms of reciprocity to address collective challenges, but such tacit understandings could break down in situations where machine intelligence is involved in human decision-making without having any normative commitments.

What is new with Artificial Intelligence? Human–agent interactions through the lens of social agency

In this article, we suggest that the study of social interactions and the development of a “sense of agency” in joint action can help determine the content of relevant explanations to be implemented in artificial systems to make them “explainable.” The introduction of automated systems, and more broadly of Artificial Intelligence (AI), into many domains has profoundly changed the nature of human activity, as well as the subjective experience that agents have of their own actions and their consequences – an experience that is commonly referred to as sense of agency. We propose to examine the empirical evidence supporting this impact of automation on individuals’ sense of agency, and hence on measures as diverse as operator performance, system explicability and acceptability. Because of some of its key characteristics, AI occupies a special status in the artificial systems landscape. We suggest that this status prompts us to reconsider human–AI interactions in the light of human–human relations. We approach the study of joint actions in human social interactions to deduce what key features are necessary for the development of a reliable sense of agency in a social context and suggest that such framework can help define what constitutes a good explanation. Finally, we propose possible directions to improve human–AI interactions and, in particular, to restore the sense of agency of human operators, improve their confidence in the decisions made by artificial agents, and increase the acceptability of such agents.

Task-irrelevant object response to action enhances the sense of agency for controlling the object in automation

The sense of agency (SoA) refers to the experience of controlling our bodies and tools. Recent automated systems require the operators to have less manual control, which decreases the SoA. This study investigated how to increase the SoA when operating automated systems, by focusing on the effect of an object's responses to operators' actions on the SoA. Participants applied brakes to a moving black circle by pressing a key, in order to stop the circle near a goal. Then, they estimated their SoA for stopping the circle. We informed them that there were automatic control trials in which the circle stopped independently of their keypress (86% of the trials). The circle's color briefly changed to white (i.e., flashed) when they pressed the key in a half of the automatic control trials. The SoA was higher with the flash than without it. However, the SoA neither increased when the circle flashed independently of a keypress nor when another object flashed due to a keypress. Furthermore, the keypress contingent object-flash did not influence the SoA when the participants controlled the circle manually. These results indicated that spatiotemporally contingent object responses to actions can enhance the SoA in automatic control situations.

The sense of agency in emerging technologies for human–computer integration: A review

Human–computer integration is an emerging area in which the boundary between humans and technology is blurred as users and computers work collaboratively and share agency to execute tasks. The sense of agency (SoA) is an experience that arises by a combination of a voluntary motor action and sensory evidence whether the corresponding body movements have somehow influenced the course of external events. The SoA is not only a key part of our experiences in daily life but also in our interaction with technology as it gives us the feeling of “I did that” as opposed to “the system did that,” thus supporting a feeling of being in control. This feeling becomes critical with human–computer integration, wherein emerging technology directly influences people’s body, their actions, and the resulting outcomes. In this review, we analyse and classify current integration technologies based on what we currently know about agency in the literature, and propose a distinction between body augmentation, action augmentation, and outcome augmentation. For each category, we describe agency considerations and markers of differentiation that illustrate a relationship between assistance level (low, high), agency delegation (human, technology), and integration type (fusion, symbiosis). We conclude with a reflection on the opportunities and challenges of integrating humans with computers, and finalise with an expanded definition of human–computer integration including agency aspects which we consider to be particularly relevant. The aim this review is to provide researchers and practitioners with guidelines to situate their work within the integration research agenda and consider the implications of any technologies on SoA, and thus overall user experience when designing future technology.

Driving With Agents: Investigating the Influences of Anthropomorphism Level and Physicality of Agents on Drivers' Perceived Control, Trust, and Driving Performance

As technological development is driven by artificial intelligence, many automotive manufacturers have integrated intelligent agents into in-vehicle information systems (IVIS) to create more meaningful interactions. One of the most important decisions in developing agents is how to embody them, because the different ways of embodying agents will significantly affect user perception and performance. This study addressed the issue by investigating the influences of agent embodiments on users in driving contexts. Through a factorial experiment (N = 116), the effects of anthropomorphism level (low vs. high) and physicality (virtual vs. physical presence) on users' trust, perceived control, and driving performance were examined. Results revealed an interaction effect between anthropomorphism level and physicality on both users' perceived control and cognitive trust. Specifically, when encountering high-level anthropomorphized agents, consumers reported lower ratings of trust toward the physically present agent than toward the virtually present one, and this interaction effect was mediated by perceived control. Although no main effects of anthropomorphism level or physicality were found, additional analyses showed that anthropomorphism level significantly improved users' cognitive trust for those unfamiliar with IVIS. No significant differences were found in terms of driving performances. These results indicate the influences of in-vehicle agents' embodiments on drivers' experience.

Hand dominance in the performance and perceptions of virtual reach control

PURPOSE

Efforts to optimize human-computer interactions are becoming increasingly prevalent, especially with virtual reality (VR) rehabilitation paradigms that utilize engaging interfaces. We hypothesized that motor and perceptional behaviors within a virtual environment are modulated uniquely through different modes of control of a hand avatar depending on limb dominance. This study investigated the effects of limb dominance on performance and concurrent changes in perceptions, such as time-based measures for intentional binding, during virtual reach-to-grasp.

METHODS

Participants (n = 16, healthy) controlled a virtual hand through their own hand motions with control adaptations in speed, noise, and automation.

RESULTS

A significant (p < 0.01) positive relationship between performance (reaching pathlength) and binding (time-interval estimation of beep-sound after grasp contact) was observed for the dominant hand. Unique changes in performance (p < 0.0001) and binding (p < 0.0001) were observed depending on handedness and which control mode was applied.

CONCLUSIONS

Developers of VR paradigms should consider limb dominance to optimize settings that facilitate better performance and perceptional engagement. Adapting VR rehabilitation for handedness may particularly benefit unilateral impairments, like hemiparesis or single-limb amputation.

Sensory Attenuation in the Auditory Modality as a Window Into Predictive Processing

Self-generated auditory input is perceived less loudly than the same sounds generated externally. The existence of this phenomenon, called Sensory Attenuation (SA), has been studied for decades and is often explained by motor-based forward models. Recent developments in the research of SA, however, challenge these models. We review the current state of knowledge regarding theoretical implications about the significance of Sensory Attenuation and its role in human behavior and functioning. Focusing on behavioral and electrophysiological results in the auditory domain, we provide an overview of the characteristics and limitations of existing SA paradigms and highlight the problem of isolating SA from other predictive mechanisms. Finally, we explore different hypotheses attempting to explain heterogeneous empirical findings, and the impact of the Predictive Coding Framework in this research area.

Transparency in Human-Machine Mutual Action

Recent advances in human-computer integration (HInt) have focused on the development of human-machine systems, where both human and machine autonomously act upon each other. However, a key challenge in designing such systems is augmenting the user’s physical abilities while maintaining their sense of self-attribution. This challenge is particularly prevalent when both human and machine are capable of acting upon each other, thereby creating a human-machine mutual action (HMMA) system. To address this challenge, we present a design framework that is based on the concept of transparency. We define transparency in HInt as the degree to which users can self-attribute an experience when machines intervene in the users’ action. Using this framework, we form a set of design guidelines and an approach for designing HMMA systems. By using transparency as our focus, we aim to provide a design approach for not only achieving human-machine fusion into a single agent, but also controlling the degrees of fusion at will. This study also highlights the effectiveness of our design approach through an analysis of existing studies that developed HMMA systems. Further development of our design approach is discussed, and future prospects for HInt and HMMA system designs are presented.

Deceleration Assistance Mitigated the Trade-off Between Sense of Agency and Driving Performance

Driving assistance technology has gained traction in recent years and is becoming more widely used in vehicles. However, drivers usually experience a reduced sense of agency when driving assistance is active even though automated assistance improves driving performance by reducing human error and ensuring quick reactions. The present study examined whether driving assistance can maintain human sense of agency during early deceleration in the face of collision risk, compared with manual deceleration. In the experimental task, participants decelerate their vehicle in a driving simulator to avoid collision with a vehicle that suddenly cut in front of them and decelerated. In the assisted condition, the system performed deceleration 100 ms after the cut-in. Participants were instructed to decelerate their vehicle and follow the vehicle that cut-in. This design ensured that the deceleration assistance applied a similar control to the vehicle as the drivers intended to, only faster and smoother. Participants rated their sense of agency and their driving performance. The results showed that drivers maintained their sense of agency and improved driving performance under driving assistance. The findings provided insights into designing driving assistance that can maintain drivers' sense of agency while improving future driving performance. It is important to establish a mode of joint-control in which the system shares the intention of human drivers and provides improved execution of control.

Embodied Cooperation to Promote Forgiving Interactions With Autonomous Machines

During every waking moment, we must engage with our environments, the people around us, the tools we use, and even our own bodies to perform actions and achieve our intentions. There is a spectrum of control that we have over our surroundings that spans the extremes from full to negligible. When the outcomes of our actions do not align with our goals, we have a tremendous capacity to displace blame and frustration on external factors while forgiving ourselves. This is especially true when we cooperate with machines; they are rarely afforded the level of forgiveness we provide our bodies and often bear much of our blame. Yet, our brain readily engages with autonomous processes in controlling our bodies to coordinate complex patterns of muscle contractions, make postural adjustments, adapt to external perturbations, among many others. This acceptance of biological autonomy may provide avenues to promote more forgiving human-machine partnerships. In this perspectives paper, we argue that striving for machine embodiment is a pathway to achieving effective and forgiving human-machine relationships. We discuss the mechanisms that help us identify ourselves and our bodies as separate from our environments and we describe their roles in achieving embodied cooperation. Using a representative selection of examples in neurally interfaced prosthetic limbs and intelligent mechatronics, we describe techniques to engage these same mechanisms when designing autonomous systems and their potential bidirectional interfaces.

Suddenly safe: Processing of conflicting feedback in assisted driving

Based on a conceptual model of operational anticipation, we propose an experimental paradigm for assisted car driving. In a pilot study, unpredicted negative outcomes, but not ambiguous intermediate feedback with positive outcomes, led to reduced feeling of safety and heightened brain activations in left anterior insular cortex attributable to error awareness.

Modified sensory feedback enhances the sense of agency during continuous body movements in virtual reality

The sense of agency refers to the feeling of control over one's own actions, and through them, the external events. This study examined the effect of modified visual feedback on the sense of agency over one's body movements using virtual reality in healthy individuals whose motor control was disturbed. Participants moved a virtual object using their right hand to trace a trajectory (Experiment 1) or a leading target (Experiment 2). Their motor control was disturbed by a delay in visual feedback (Experiment 1) or a 1-kg weight attached to their wrist (Experiment 2). In the offset conditions, the virtual object was presented at the median point between the desired position and the participants' actual hand position. In both experiments, participants reported improved sense of agency in the offset condition compared to the aligned condition where the visual feedback reflected their actual body movements, despite their motion being less precise in the offset condition. The results show that sense of agency can be enhanced by modifying feedback to motor tasks according to the goal of the task, even when visual feedback is discrepant from the actual body movements. The present study sheds light on the possibility of artificially enhancing body agency to improve voluntary motor control.

Explicit Sense of Agency in an Automatic Control Situation: Effects of Goal-Directed Action and the Gradual Emergence of Outcome

Sense of agency (SoA), or the subjective feeling that "I am the agent controlling the object," is essential for learning and enjoying object manipulation. Recently developed automatic control systems, such as the cruise control systems in autonomous vehicles, require less manual control from the manipulators. It has to date been impossible to completely relieve operators of the need for manual control in many automatic control systems developed for tool-using situations. Therefore, it is important to examine how to maintain SoA (illusorily) during an automatic control situation. We investigated the effects of two typical characteristics of everyday tool-use situations on SoA when braking a moving object with a keypress. These characteristics included the presence of a goal (e.g., in driving situations, the driver steps on the brake pedal to stop the car at an expected position) and the gradual emergence of the outcome (e.g., the driver steps on the brake pedal and the car usually slows down first and then stops). We conducted an experiment in which participants stopped a moving object and then rated their SoA for stopping the object. Participants were explicitly informed that the object would sometimes stop independently of their keypress. Results showed that both characteristics decreased SoA in the manual control situation but increased SoA in the automatic control situation. Thus, these characteristics could be useful for maintaining a sense of agency in automatic control situations.

The Role of Machine Learning in Spine Surgery: The Future Is Now

The recent influx of machine learning centered investigations in the spine surgery literature has led to increased enthusiasm as to the prospect of using artificial intelligence to create clinical decision support tools, optimize postoperative outcomes, and improve technologies used in the operating room. However, the methodology underlying machine learning in spine research is often overlooked as the subject matter is quite novel and may be foreign to practicing spine surgeons. Improper application of machine learning is a significant bioethics challenge, given the potential consequences of over- or underestimating the results of such studies for clinical decision-making processes. Proper peer review of these publications requires a baseline familiarity of the language associated with machine learning, and how it differs from classical statistical analyses. This narrative review first introduces the overall field of machine learning and its role in artificial intelligence, and defines basic terminology. In addition, common modalities for applying machine learning, including classification and regression decision trees, support vector machines, and artificial neural networks are examined in the context of examples gathered from the spine literature. Lastly, the ethical challenges associated with adapting machine learning for research related to patient care, as well as future perspectives on the potential use of machine learning in spine surgery, are discussed specifically.