The study arrives at a moment of tension in the deployment of self-driving technology. On the technical side, the numbers are impressive. Drawing on prior real-world evidence, the authors note that across 56.7 million rider-only miles, autonomous systems showed crash-rate reductions of roughly 79 percent for injury-reported crashes, 81 percent for airbag-deployment crashes, and 85 percent for suspected serious-injury crashes compared with human benchmarks. Bodily-injury insurance claims in fully driverless operations fell by as much as 92 to 93 percent, and police-reported injury crashes dropped by about 55 percent in some analyses. Since human error contributes to roughly 94 to 95 percent of road collisions, these gains represent a genuine transformation in road safety.

Yet the study argues that functional safety and perceived safety are fundamentally distinct constructs, and that conflating them is a mistake with real consequences for public acceptance. The researchers, led by Anu Masso of Tallinn University of Technology together with colleagues in Hong Kong and Denmark, designed a semi-experimental online study in which participants viewed paired traffic scenarios presented under both autonomous and non-autonomous conditions. The scenarios deliberately featured a wide range of social diversity cues, including gender, age, race operationalized through skin-tone variation, income, disability, and micromobility, grounded in theories of super-diversity and intersectionality. Areas of interest were matched in size and position across paired images to rule out low-level visual saliency as a confound.

Eye tracking was conducted through the RealEye platform, which uses webcam-based AI gaze prediction with an accuracy of roughly 110 pixels. Each trial began with a 40-point calibration and a 9-point validation check, and participants whose calibration errors exceeded 2.5 degrees of visual angle or who failed attention checks were excluded. The primary measure was fixation total time, the cumulative dwell time spent on each social diversity area of interest, chosen for its robustness in webcam-based settings and its interpretability as sustained attentional engagement. Revisit counts served as secondary validation. The team then used Bayesian multilevel regression models, estimated with four Markov Chain Monte Carlo chains in the brms package, to link attention patterns with self-reported safety and security evaluations while accounting for individual and city-level variation.

The first major finding concerns the structure of visual attention itself. Gaze patterns toward social groups proved to be highly organized and remarkably stable across both autonomous and non-autonomous scenarios. Categories such as age, race, and income consistently attracted higher levels of attention in both contexts, while disability-related and COVID-related cues remained comparatively backgrounded. Automation modestly increased overall attention to social cues without fundamentally reshaping these hierarchies. The estimated association between attention indices in the two contexts was moderate, with a beta of approximately 0.66 and a 95 percent confidence interval spanning 0.44 to 0.88, indicating that attentional structure is partially stable but also partially reconfigured by mobility context. In other words, autonomous mobility amplifies existing perceptual salience hierarchies rather than creating new ones.

The surprise came when the researchers connected those gaze patterns to feelings of safety. At the individual level, correlations between visual attention indices and perceived safety or security were weak, with absolute correlation values below 0.2, and scatter plots revealed no clear linear associations. In the final Bayesian models, the key predictor was the difference in attention between autonomous and non-autonomous contexts, and its posterior estimates were centered near zero with credible intervals overlapping zero for both outcomes. Once trust, social positioning, and regional context were accounted for, shifts in visual attention simply did not translate into changes in experienced safety or normative security. Being perceptually alert to social difference, it turns out, is not the same as feeling protected by it.

What did predict perceived safety? Trust emerged as the dominant force. Both general trust in automation, measured across SAE automation Levels 0 through 5, and situational trust in autonomous transport applications showed consistently positive posterior estimates for experienced safety. Notably, trust remained relatively stable across automation levels 0 to 4, with mean scores between 3.40 and 3.49, but dropped to 3.10 for Level 5 full automation, signaling persistent hesitation toward fully driverless systems even among people otherwise comfortable with semi-automated technology. Trust also varied by who was being transported: respondents expressed the highest trust in autonomous vehicles carrying people from rural to urban areas and the lowest in vehicles carrying elderly people to parks, suggesting that acceptance depends on the perceived vulnerability and social status of passengers.

Security values followed a different explanatory logic altogether. Unlike perceived safety, security as a normative principle was less strongly associated with trust and more strongly shaped by social position and cultural context. Minority status, measured through self-identification as belonging to a national minority group, was the most robust individual-level predictor, with minority respondents systematically assigning lower importance to security values and reporting lower perceived safety even after controlling for trust and attention. Regional effects were substantial in both models, with patterns more pronounced for security values than for perceived safety, reinforcing the interpretation that security evaluations are deeply anchored in cultural and institutional environments. Higher perceived safety was generally reported by respondents with higher education and by those living in European and East and Southeast Asian regions, while younger respondents aged 18 to 29 and those in the Americas or Sub-Saharan Africa reported lower levels.

The broader survey results add texture to this picture. Nearly 30 percent of respondents felt that transport safety in their city had remained unchanged over the previous five years, while roughly 55 percent perceived an increase and about 16 percent a decrease, indicating that even perceptions of objective safety trends are filtered through local experience. Meanwhile, security as a public value was overwhelmingly salient: 54.7 percent rated it as very important in the development of AI-based urban technologies, and only 13 percent considered it insignificant. These patterns suggest that while people care intensely about security in the abstract, their concrete feelings of safety in automated mobility are constructed through trust, social identity, and institutional context rather than through what their eyes happen to land on.

The implications for the autonomous vehicle industry and for AI ethics are significant. A widespread assumption holds that making diverse social groups visible and recognizable within mobility environments is sufficient to foster inclusion, trust, and safety. This study challenges that assumption directly. Visual representation may be necessary for functional safety, ensuring that systems reliably detect vulnerable road users, but it does not guarantee psychological reassurance or normative legitimacy. The authors call for complementing safety-by-design with socio-psychological safety-by-design, in which trust calibration, cultural legibility, and social recognition extend beyond what systems detect or display. The study also acknowledges limitations, including the lower spatial precision of webcam-based eye tracking compared with laboratory systems, the possibility that visual stimuli unintentionally encoded stereotypes, and small sample sizes in some cities, which the authors address through partial pooling in their Bayesian models. Future research, they suggest, should combine laboratory-grade eye tracking, in situ mobility experiences, and qualitative approaches to uncover the mediating mechanisms, such as institutional trust and prior experience with automation, that connect perception to judgment. For now, the message is clear: legitimate autonomous mobility will not be won by crash statistics or visible diversity alone, but by the social and institutional conditions under which people come to feel genuinely safe.

Subject of Research: Cross-cultural eye-tracking research on perceived safety and trust in autonomous vehicle environments

Article Title: What shapes perceived safety in autonomous mobility? Cross-cultural evidence from large-scale online eye tracking

Article References: Masso, A., Pietarinen, A.-V., Ibrahimi, M., & Ben Yahia, S. (2026). What shapes perceived safety in autonomous mobility? Cross-cultural evidence from large-scale online eye tracking. AI & SOCIETY. https://doi.org/10.1007/s00146-026-03278-1

Image Credits: AI Generated

DOI: 10.1007/s00146-026-03278-1

Keywords: autonomous vehicles, eye tracking, perceived safety, trust in automation, AI ethics, cross-cultural study, social diversity, Bayesian multilevel models, data justice, minority status, human-machine interaction, urban mobility

The study addresses a problem that has grown more urgent as autonomy levels in commercial and experimental surgical systems have climbed. A 2024 systematic review of FDA-cleared surgical robots documented a wide spectrum of autonomy, from teleoperated tools that merely filter tremor to systems executing defined subtasks without continuous guidance. Meanwhile, laboratory demonstrations have grown bolder, including deep learning-based autonomous retinal vein cannulation in ex vivo porcine eyes reported in Science Robotics in 2025. Yet clinicians and ethicists have repeatedly cautioned that a human must remain meaningfully in the loop, and research on interruptions in healthcare shows that poorly timed requests for attention can themselves become a safety hazard. The Nagoya group’s question was therefore not whether robots should ask for help, but when.

Technically, the framework supports three distinct interaction modalities for soft tissue resection, the task of cutting away target tissue. In the Autonomous modality, the robot executes the resection without any feedback requests, relying entirely on its internal perception and planning. In the Supervised modality, the robot pauses before every critical action and requires explicit confirmation from the human operator, a design that maximizes oversight but also maximizes the number of interruptions. The third option, which the authors call the Selective modality, is the novel contribution: the robot monitors its own internal uncertainty estimates and initiates communication with the operator only when those estimates cross a predefined threshold. When the system is confident about the cutting path, it proceeds silently; when it is not, it asks.

The machinery behind that confidence check draws on a well-established body of work in Bayesian deep learning and uncertainty estimation. Techniques such as deep ensembles, in which multiple neural networks are trained independently and their disagreement is treated as a proxy for uncertainty, and the decomposition of predictive uncertainty into aleatoric and epistemic components, as formalized by Kendall and Gal, give modern perception systems a way to know what they do not know. The team has previously applied related ideas to surgical workflow recognition and to monocular depth estimation for surgical scenes, and a companion study explored large language model-based detection of ambiguity in natural language instructions given to collaborative surgical robots. The Selective modality effectively converts those uncertainty signals into a communication policy.

To evaluate the approach, the researchers ran a preliminary user study in which participants performed resection of phantom tissue, a standard surrogate that mimics the mechanical properties of soft tissue without the ethical and regulatory complexity of real patients. Each interaction modality was assessed along two complementary axes. Quantitative reliability was measured through cutting path error, which captures how closely the executed cut matched the intended trajectory, and task completion time. Qualitative measures targeted the human side of the collaboration: the NASA Task Load Index quantified mental demand and overall workload, while validated trust and distrust questionnaires, building on foundational scales for trust in automated systems and more recent instruments for calibrating trust in artificial intelligence, captured how participants perceived the system’s reliability.

The results, though preliminary, point in a consistent direction. Compared with the Supervised modality, the Selective modality showed a trend toward lower mean path error and shorter mean completion time, suggesting that removing the requirement to confirm every single action did not come at the cost of precision and may actually have helped by keeping the operator’s attention focused where it mattered. Compared with the Autonomous modality, the Selective condition preserved the safety benefit of human involvement at the moments of greatest risk, since queries were concentrated precisely where the robot’s internal models were least certain.

The trust findings are arguably the most intriguing part of the study. Participants in the Selective condition reported lower mean distrust scores and lower mental demand than in the comparison conditions, while positive trust scores remained similar to those observed elsewhere. In other words, asking for help sparingly did not erode confidence in the robot; if anything, it reduced the suspicion and cognitive strain associated with a machine that either never checks in or never stops checking in. This aligns with a broader theme in human-robot interaction research, which suggests that appropriately timed communication can calibrate a human operator’s mental model of what an automated system is doing, rather than simply flattering it with reassurance.

The authors are careful about the limits of what they can claim. The study is explicitly labeled preliminary, the sample was small, and the observed differences in path error, completion time, and questionnaire scores are trends rather than statistically confirmed effects. The team states plainly that a larger study is required to determine whether the patterns they observed represent reliable phenomena. Phantom tissue, however realistic, also differs from living anatomy in ways that matter for bleeding, deformation, and visual appearance, so extending the evaluation toward more clinically representative scenarios is an obvious next step. The experimental protocols were approved by the Ethical Research Committee of Nagoya University, and informed consent was obtained from all participants.

Even so, the framework’s design philosophy has implications well beyond one lab bench. The work was supported by the Japan Science and Technology Agency’s CREST program, including the AIP Challenge Program, and by JSPS KAKENHI grants, reflecting a sustained national investment in human-centered automation. The idea of uncertainty-gated communication is modular: it does not dictate what the robot’s perception system must be, only that whatever confidence estimates it produces should be translated into a disciplined policy about when to involve the human. That makes the approach potentially compatible with the growing family of autonomous surgical assistance functions, from exposure maximization during dissection and cautery to tool exchange and workflow tracking, each of which could inherit the same selective interaction layer.

For surgeons, the near-term significance is a possible answer to a daily ergonomic complaint: automation that either demands constant babysitting or offers none at all forces an uncomfortable choice. A system that interrupts only when its internal uncertainty crosses a threshold promises to reserve human attention for the decisions that genuinely need it, while giving the operator a continuous, implicit signal about the machine’s state, because silence itself becomes informative. For patients, the longer-term promise is a class of surgical assistants that combine machine precision with human judgment in a way that is measurable rather than rhetorical. The Nagoya team’s data are early, but they sketch a credible engineering route toward surgical robots that are not just capable of autonomy, but judicious about exercising it, and that may prove to be the quality on which the trust of operating rooms ultimately turns.

Subject of Research: Uncertainty-triggered human-robot interaction strategies for reliable and trustworthy collaborative robotic surgical assistance

Article Title: Assessing selective interaction for reliable and trustworthy robotic surgical assistance

Article References: Colan, J., Davila, A., Yamada, Y., Misawa, K., & Hasegawa, Y. (2026). Assessing selective interaction for reliable and trustworthy robotic surgical assistance. International Journal of Computer Assisted Radiology and Surgery. https://doi.org/10.1007/s11548-026-03799-6

Image Credits: AI Generated

DOI: 10.1007/s11548-026-03799-6

Keywords: surgical robotics, human-robot collaboration, uncertainty estimation, selective interaction, soft tissue resection, NASA-TLX, trust in automation, autonomous surgery, Bayesian deep learning, phantom tissue, workload, Nagoya University