The study, authored by Brady Kennedy of Columbia University’s Department of Sociology, Cristian Capotescu of Columbia’s Trust Collaboratory, and Jennifer Laird of Lehman College, challenges the prevailing assumption that trust in smart city technologies can be engineered through better privacy protections, more reliable sensors and clearer communication. Drawing on a thematic synthesis of sociological research on trust alongside the literature on public participation, the authors contend that trust is not an attitude stored in the minds of individuals and measured by surveys, but a property of ongoing social relations between residents and the experts and institutions behind these initiatives.

Smart city initiatives typically combine several layers of technology: distributed sensors and connected devices grouped under the heading of the Internet of Things, which collect data from the built environment; information and communication technologies and cloud computing, which transmit and process that data at scale; and increasingly, artificial intelligence and algorithmic decision-making, which translate the resulting data into classifications and predictions for operational use. Stated aims range from optimizing public services and improving environmental sustainability to making infrastructure more accessible and cities safer and more livable.

Much existing research on trust in smart cities operates within a risk-benefit logic. Scholars have produced testable models simulating how mistrust hinders deployment, measured public attitudes through surveys, and proposed lists of trust characteristics such as privacy, reliability and fairness that designers can integrate into systems. The assumption is that as initiatives follow prescribed security parameters, privacy standards and governance norms, and communicate those choices transparently, public trust will follow. Kennedy and colleagues call this a technology-centric view that treats trust as flowing in one direction, with the public asked to extend it according to how experts adjust technical features.

In contrast, recent sociological scholarship, particularly from the field of science, technology and society, understands trust as inherently relational. In this account, trust emerges from interactions between trustors, the people doing the trusting; trustees, those in whom trust is placed; and trust objects, the specific focus of the relation at a given time and place. A smart city initiative may be the object of a trust relation, but it is never the trustee: residents extend or withhold trust with respect to the agency, vendor or partnership that designs and operates the system. Crucially, because uncertainty is what makes trust necessary in the first place, its conditions can never be fully engineered away through technical decisions.

The authors further draw on work describing trust as a verb, an active practice of trusting. Trusting is a skillful evaluative process in which residents scrutinize information and conduct rather than placing blind faith in experts. This carries a demand for both sides: the trustor must do the work of assessment, and the trustee must render itself answerable to that scrutiny. Because the trustee in these relations is an institution whose authority is itself at issue, answerability requires that decisions, and the power behind them, be made contestable. Notably, the authors argue that mistrust is not a failure of trust-building but internal to the same relational ties, playing an important role in promoting accountability and public engagement.

From this foundation, the study develops a typology of three participatory roles: the public as co-accessors, co-collectors and co-designers. Each role names the position residents occupy relative to smart city experts, and, because a relation cannot be specified from one side alone, also names what experts must place at stake. As co-accessors, residents shape, monitor and contest how data is gathered and used, while institutions accept the vulnerability of scrutiny. Examples range from static open data dashboards such as London’s DataStore to deeper arrangements like Barcelona’s New Data Deal, which re-envisions collected data as a public commons, and Manchester’s data governance practices, where residents and policymakers negotiate data gathering within structured engagements.

As co-collectors, residents directly produce data for smart city systems, through passive smartphone sensing or active crowdsourcing, while experts accept dependence on evidence they did not themselves produce. The MiraMap platform in Torino, Italy, allows residents to report and track urban problems, research showing it increased trust in government and technology alike. At the furthest extent, residents deploy their own sensors through systems such as HabitatMap AirSense and AirBox to generate environmental evidence independent of official accounts, converting mistrust into an instrument of accountability. As co-designers, finally, residents participate in shaping what is built, with experts relinquishing exclusive control. Iterative workshop approaches that visibly incorporate resident feedback have been shown in follow-up surveys to increase community trust in both the technology and the researchers, while game-based and virtual reality formats help include elderly and non-technical participants.

The framework carries pointed lessons for practitioners. Tokenism and consultation that leave residents feeling passively managed are not failures of technique but interactions in which no reciprocity was established; the mistrust that follows is a proportionate response, as illustrated by the urban data trust proposed for Sidewalk Toronto, constituted by the operator on the operator’s own terms, which failed to generate the mutuality on which trusting depends. The authors also caution that participatory approaches do not encounter a ready-made public but bring one into being, and that formats demanding time, mobility and technical confidence tend to reproduce existing inequalities, underrepresenting marginalized groups. Compensation and trusted community intermediaries can help widen who participates.

Ultimately, the authors argue, smart city professionals should stop treating participation as an instrument for raising approval scores and start treating it as the intentional structuring of reciprocal interaction in which trust, and mistrust, can both operate productively. There is no single modality for building trust, and each initiative will pose context-specific challenges. But without a relational understanding in which experts make themselves genuinely answerable, they warn, even well-intentioned future initiatives will face the same public backlash that has already derailed smart city projects from Toronto to San Diego.

Subject of Research: Trust-building through public participation in smart city initiatives

Article Title: Reframing participatory approaches to the smart city as opportunities for trust-building

Article References: Kennedy, B., Capotescu, C., & Laird, J. (2026). Reframing participatory approaches to the smart city as opportunities for trust-building. Discover Cities, 3(1), Article 177. https://doi.org/10.1007/s44327-026-00352-7

Image Credits: AI Generated

DOI: 10.1007/s44327-026-00352-7

Keywords: smart cities, public trust, public participation, sociology of trust, STS, expertise, data governance, participatory design, citizen sensing, urban technology, co-design, mistrust

The scientific foundation for this work rests on well-established theories of expertise. The information-reduction hypothesis holds that experts become efficient by discarding irrelevant visual information and concentrating on the crucial elements of a task. The theory of long-term working memory suggests that mastery extends processing capacity by building retrieval structures that let experts access vast stores of knowledge quickly. Meanwhile, the holistic model of image perception, developed from mammography research, proposes that experts gather information from broad and peripheral areas of the visual field, effectively widening their useful field of view. Decades of eye-tracking studies have backed these ideas: as professionals gain experience, they visit objects less frequently, spend less time viewing them, avoid distractors more reliably, and show evidence of an expanding visual span.

What makes the new review distinctive is its focus on turning those expert differences into training tools. The authors systematically searched the Scopus database following PRISMA guidelines, screening 79 initially identified articles down to a final sample of eight studies that either examined expertise-dependent differences in eye movements or tested the transfer of skills through gaze patterns. Their conceptual framework divided the field into three elements: the skill-transfer methods and cognitive phenomena involved, the macro application areas where eye movements serve as training tools, and the eye-tracking metrics used to measure both expertise and learning. The framework captures how techniques such as eye movement modeling examples, or EMMEs, record an expert’s gaze as the expert performs a task, then present the playback to learners alongside the expert’s verbal narration.

EMMEs are theoretically rooted in observational learning and in the cognitive theory of multimedia learning, and they draw on a striking discovery in cognitive neuroscience: the brain’s mirror system activates when a person watches another perform an action, simulating that action internally. Research on expert dancers has shown that this mirroring process can help integrate observed actions into the observer’s own behavioral repertoire. In practical studies, the technique has delivered real gains. Novice aircraft inspectors trained with expert gaze displays detected more faults during search tasks, and EMME-trained inspectors of circuit boards showed improved fault detection. Even programmers using expert gaze cues debugged software more quickly, suggesting the approach extends well beyond medicine and industry.

Medicine, and especially surgery, has become the proving ground for gaze-based training. In a randomized controlled trial, novice surgeons trained to follow expert-like gaze strategies on a laparoscopic simulator developed more target-locking fixations, completed tasks faster, and made fewer errors than peers who learned by discovery alone. The advantage became even more pronounced when participants had to multitask, hinting that trained gaze frees cognitive resources for other demands. Related work on quiet eye training, a technique that extends the final fixation before a critical movement, showed that trainees who practiced knot-tying with gaze training maintained their performance under heightened anxiety while traditionally trained peers faltered. Collaborative systems that displayed a supervisor’s live gaze to trainees reduced completion times and errors, and studies of visual guidance during laparoscopic tasks found better trainee performance when experts’ point of gaze was made visible.

The applications reach far beyond the operating room. In a maritime operation simulator, researchers built expert-derived attention maps that told trainees exactly where to focus during heavy lifting operations; the briefed group showed superior visual focus compared with a control group told only about the risks. In construction, eye-tracking studies revealed that workers who scanned the workplace more broadly recognized a higher proportion of hazards, and that personalized feedback based on eye movement data improved both search patterns and hazard recognition. In aviation, researchers comparing helicopter pilots of different experience levels during landing simulation found that veterans relied more heavily on cockpit instruments while novices looked out the window, and that eye-tracking feedback improved the transfer of skills from simulator to real flight.

Not every finding supported the enthusiasm. Studies of EMMEs in school settings produced mixed results. One experiment found that students who watched a model’s gaze replay while reading illustrated texts integrated verbal and graphical information more effectively, and that weaker readers benefited most. But another pair of experiments on procedural problem-solving in geometry found no significant advantage from displaying a model’s eye movements, and in one case the modeled gaze actually slowed transfer problem-solving. The authors of the review conclude that the nature of the task is a critical moderator: gaze modeling appears most effective for non-procedural tasks such as classification and strategy learning, where the expert simply observes the material, and least effective for procedural tasks requiring direct interaction with on-screen objects, which already capture attention naturally.

The review also flags subtler moderators. Prior knowledge matters, consistent with the expertise reversal effect, which holds that extra instructional guidance can burden learners who already know enough to proceed without it. The presence of verbal explanations alongside gaze overlays can either enhance learning by revealing the expert’s covert cognitive processes or, for perceptually simple tasks, overload the learner with redundant information. Interestingly, one study of medical image diagnosis found that experienced experts benefited from gaze modeling even more than novices did, improving diagnostic performance, scanning efficiency, and their ability to adapt skills to unfamiliar visualizations. This suggests that eye movement modeling examples can promote adaptive expertise, helping even seasoned professionals confront the evolving technologies that constantly reshape their fields.

The authors are candid about the limits of their evidence. Only eight studies met the strict inclusion criteria, and sample sizes were small and methodologically heterogeneous, so the performance advantages of gaze training should be interpreted with caution and cannot yet be generalized. Eye trackers measure only foveal vision, missing the covert shifts of attention that let people process information in peripheral and parafoveal regions, and the cost of research-grade systems, ranging from thousands to tens of thousands of euros, remains a barrier to widespread adoption. The exclusive reliance on Scopus may also have omitted relevant studies indexed elsewhere. Still, the review’s core conclusion stands: experts are more focused, organized, and deliberate in their visual behavior than novices, gaze-trained individuals show measurable performance advantages in the tasks examined, and rendering the invisible perceptual strategies of expertise visible may be one of the most promising frontiers in professional training, with implications for medicine, aviation, industry, sport, and education alike.

Subject of Research: A systematic review of skills-transfer and gaze strategies studied through eye-tracking across professional domains

Article Title: Skills-Transfer and Gaze Strategies Studied by Eye-Tracking: A Systematic Review

Article References: Lupia, E., Bortolotti, A., & Palumbo, R. (2026). Skills-Transfer and Gaze Strategies Studied by Eye-Tracking: A Systematic Review. Trends in Psychology. https://doi.org/10.1007/s43076-026-00535-6

Image Credits: AI Generated

DOI: 10.1007/s43076-026-00535-6

Keywords: eye-tracking, expertise, skill transfer, gaze training, quiet eye, eye movement modeling examples, visual attention, surgical training, human performance, systematic review, cognitive science, training methods