Across domains, organizations and society are facing a trust deficit (Twenge, Campbell, & Carter, 2014). This is problematic, as trust is important to a variety of critical organizational outcomes, such as perceived task performance, team satisfaction, relationship commitment, and stress mitigation (Costa, Roe, & Taillieu, 2001), and has been cited as a motivator for cooperation and knowledge transfer due to its capacity to reduce fear and risk of exploitation (Chen et al., 1998; Fleig-Palmer & Schoorman, 2011; Irwin & Berigan, 2013; Yamagishi & Sato, 1986), and a key component of collaboration. As organizations increasingly rely upon collaboration for achieving important outcomes, it is of critical importance that organizations understand how to not only develop interpersonal trust in collaborative partnerships to facilitate these positive outcomes, but also the way in which interpersonal trust is broken and can be repaired when problems inevitably arise. Though research has begun to investigate trust violation and trust repair, relatively little is known about trust development, violation, and repair as a process that unfolds over time. This is problematic, as cross-sectional studies fail to capture change, both in terms of how trust itself changes as well as how the effect of a violation or the utility of a repair strategy may be weaker or stronger in the long-term than the short-term. Thus, findings from a single point in time may result in different conclusions and recommendations than those that would result from long-term investigation. Therefore, this study examines how interpersonal trust patterns unfold within individuals, and how these patterns differ between individuals depending on the type of violation and the repair strategy employed. An experimental study using discontinuous growth modeling to examine intraindividual and interindividual differences in trust processes found that generally, trust was negatively impacted more after an intentional ("will do") violation as compared to a competence ("can do") violation, such that it had a greater impact on character assessments than a competence violation and also damaged perceptions of ability as much as a competence violation. These negative impacts carried over into trust restoration, which was significantly slower after an intentional violation than a competence violation. Furthermore, study findings suggest that after an intentional violation, trust restored more quickly when surveillance was implemented than when compensation was offered. Though the opposite did not hold true for a competence violation, the findings did approach significance. Drawing from these findings, implications and future research recommendations are discussed.

Trustworthiness is a roadblock in mass adoption of artificial intelligence (AI) in medicine. This thesis developed a framework to explore the trustworthiness as it applies to AI in medicine with respect to common stakeholders in medical device development. Within this framework the element of explainability of AI models was explored by evaluating explainable AI (XAI) methods. In current literature a litany of XAI methods are available that provide a variety of insights into the learning and function of AI models. XAI methods provide a human readable output for the AI’s learning process. These XAI methods tend to be bespoke and provide very subjective outputs with varying degrees of quality. Currently, there are no metrics or methods of objectively evaluating XAI outputs against outputs from different types of XAI methods. This thesis presents a set of constituent elements (similarity, stability and novelty) to explore the concept of explainability and then presents a series of metrics to evaluate those constituent elements. Thus providing a repeatable and testable framework to evaluate XAI methods and their generated explanations. This is accomplished using subject matter expert (SME) annotated ECG signals (time-series signals) represented as images to AI models and XAI methods. A small subset from all available XAI methods, Vanilla Saliency, SmoothGrad, GradCAM and GradCAM++ were used to generate XAI outputs for a VGG-16 based deep learning classification model. The framework provides insights about XAI method generated explanations for the AI and how closely that learning corresponds to SME decision making. It also objectively evaluates how closely explanations generated by any XAI method resemble outputs from other XAI methods. Lastly, the framework provides insights about possible novel learning done by the deep learning model beyond what was identified by the SMEs in their decision making. / Thesis / Master of Applied Science (MASc) / The goal of this thesis was to develop a framework of how trustworthiness can be improved for a variety of stakeholders in the use of AI in medical applications. Trust was broken down into basic elements (Explainability, Verifiability, Fairness & Ro- bustness) and ’Explainability’ was further explored. This was done by determining how explainability (offered by XAI methods) can address the needs (Accuracy, Safety, and Performance) of stakeholders and how those needs can be evaluated. Methods of comparison (similarity, stability, and novelty) were developed that allow an objective evaluation of the explanations from various XAI methods using repeatable metrics (Jaccard, Hamming, Pearson Correlation, and TF-IDF). Combining the results of these measurements into the framework of trust, work towards improving AI trust- worthiness and provides a way to evaluate and compare the utility of explanations.

The goal of this study is to determine if trust in autonomous vehicles is affected by whether the vehicle is being operated in a closed or open system. A PRQF survey method was used to complete this study. The survey contained items to assess pedestrian behavior, personal innovativeness, and receptivity to autonomous vehicles. Scenario questions were also utilized to determine differences in the trust of automated vehicles in open and closed settings. The results from this study indicated increased pedestrian receptivity scores for the closed system (M=14.11, SD=3.78), compared to the open system (M=13.70, SD=3.90). Average trust scores were also increased for the closed system (M=4.68, SD=1.82) compared to the open system (M=4.56, SD=1.85). These results were used to conclude that trust and receptivity of autonomous vehicles were increased for closed systems.

Thesis advisor: Raquel Muñiz / This qualitative case study explores the role that trust plays between the superintendent and the teacher union leader of a public school district in the Northeast United States during the COVID-19 pandemic using the framework of interpersonal trust-building (Zand, 1972). Further, it uses the five facets of trust (Tschannen-Moran, 2001) to identify the leadership practices that have the greatest impact on perceptions of trust in this relationship. Specifically, this study addresses the following research question: How, if at all, does trust influence the relationships and practices of educational stakeholders during times of crisis? Based on semi-structured interviews, document reviews, and observations, findings support previous research indicating that trust develops only with the benefit of time and, once established, allows for more direct communication and more efficient and collaborative problem-solving. Data also indicate that the facet of benevolence exerts the greatest impact on perceptions of trust in the superintendent’s and teacher union leader’s working relationship. Finally, the accumulation of shared experiences over time help develop a shared sense of identity between the superintendent and teacher union leader, resulting in stronger perceptions of trust and a greater sense of shared purpose. This shared sense of identity may also serve as a proxy for time, allowing parties to make assumptions about the other’s future behavior based on perceived group memberships, thereby jump-starting the development of trust in the relationship. Recommendations include purposefully demonstrating benevolent behaviors in order to more effectively develop trust in a relationship and, whenever possible, communicating a shared sense of identity based on common values and beliefs. These findings have implications for district and school leaders who want to more intentionally establish trusting relationships and can inform the preparation, induction, and learning of district leaders. / Thesis (EdD) — Boston College, 2022. / Submitted to: Boston College. Lynch School of Education. / Discipline: Educational Leadership and Higher Education.

At the age of eighteen, when I was approaching the end of my apprenticeship as a cabinetmaker, I made my first self designed piece of furniture. The master cabinetmaker or the client determined the form of most of the furniture made in our shop, and I seldom liked it. I did not even like the wood we used for the best pieces: walnut. I chose light colored ash for my bed and cupboard, and I made them so they looked good on all sides, with the same wood and the same careful work back and front. I disregarded the usual practice of expanding less time and care on the back because no one ever sees it anyway. At long last I was able to round off the edges only slightly without being corrected, running the sandpaper swiftly and lightly over the edges to soften their sharpness without losing the elegance and fineness of the lines. I barely touched the corners where the three edges met. I fitted the door of the cupboard into the frame at the front with a maximum of precision so that it closed almost hermetically, with a gentle frictional resistance and a barely audible sound of escaping air. It felt good working on this cupboard. Making the precisely fitting joints and exact shapes to form a whole, a complete object that corresponds to my inner vision, triggering in me a sense of intense concentration, and the finished piece of furniture added freshness to my environment. Peter Zumthor, Thinking Architecture / Master of Architecture

The purpose of this study was to use quantitative data to answer the overarching question, what behaviors develop mutual trust, and what is its association to job satisfaction? This study expanded on the research of Tschannen-Moran, considering mutual trust and job satisfaction. Using quantitative research, the researcher considered the five constructs of trust originally identified by Hoy and Tschannen-Moran to determine what behaviors are needed on the part of the principals and teachers to develop mutual trust and if there is an association between mutual trust and job satisfaction. The quantitative data were collected from four schools in a medium-sized rural school division in Virginia. Survey questions were developed with the five constructs of trust from Tschannen-Moran's research as the foundation and provided information on principals' and teachers' behaviors. Most questions were formatted to use a 5-point Likert scale; however, two open-ended questions provided more specific information on behaviors needed to develop mutual trust. Through analysis of the data, the researcher found that the development of trust is primarily the result of the behaviors of the principals. Further, there are differences between the perceptions of principals and teachers regarding behaviors needed to develop trust. This information is helpful for future educators, both principals and teachers. As part of an individual's preparation for a career in education, it is beneficial for them to have some understanding of how to develop a trusting relationship in a school between a teacher and principal. Current administrators need to understand if there is an overall culture of trust in the school they lead, and if not, where the gap in trust is. / Ed. D.

The rapid growth of the Internet has resulted in the desperate need for alternative ways to keep electronic transactions secure while at the same time allowing entities that do not know each other to interact. This has, in turn, led to a wide area of interest in the issues of trust and trust modeling to be used by machines. A large amount of work has already been undertaken in this area in an attempt to transfer the trust and interaction decision making processes onto the machine. However this work has taken a number of different approaches with little to no correlation between various models and no standard set of criteria was even proposed that can be used to evaluate the value of such models. The proposed research chooses to use a detailed literature survey to investigate the current models in existence. This investigation focuses on identifying criteria that are required by trust models. These criteria are grouped into four categories that represent four important concepts to be implemented in some manner by trust models: trust representation, initial trust, trust update and trust evaluation. The process of identifying these criteria has led to a second problem. The trust evaluation process is a detailed undertaking requiring a high processing overhead. This process can either result in a value that allows an agent to trust another to a certain extent or in a distrust value that results in termination of the interaction. The evaluation process required to obtain the distrust value is just as process intensive as the one resulting in determining a level of trust and the constraints that will be placed on an interaction. This raises the question: How do we simplify the trust evaluation process for agents that have a high probability of resulting in a distrust value? This research solves this problem by adding a fifth category to the criteria already identified; namely: prejudice filters. These filters have been identified by the literature study and are tested by means of a prototype implementation that uses a specific scenario in order to test two simulation case studies. / Dissertation (MSc)--University of Pretoria, 2008. / Computer Science / unrestricted

Trust is the cornerstone of success in any relationship between two or more parties. Generally, we do not socialize, seek advice, consult, cooperate, buy or sell goods and services from/to others unless we establish some level of mutual trust between interacting parties. When e-commerce was merging infancy, the concept of trusting an entity in a virtual world was a huge obstacle. Gradually, increasingly-sophisticated, largely generic reputation scoring and management systems were embedded into the evolving marketplaces. Current technologies to include cloud computing, social networking, and mobile applications, coupled with the explosion in storage capacity and processing power, are evolving large-scale global marketplaces for a wide variety of resources and services, such as Amazon.com, BitTorrent, WebEx and Skype. In such marketplaces, user entities, or users for short; namely, consumers, providers and brokers, are largely autonomous with vastly diverse requirements, capabilities and trust profiles. Users' requirements may include service quality levels, cost, ease of use, etc. Users' capabilities may include assets owned or outsourced. Trustors' profiles may include what is promised and commitments to keep these promises. In such a large-scale heterogeneous marketplace, the trustworthy interactions and transactions in services and resources constitute a challenging endeavor. Currently, solving such issues generally adopts a "one-size fits all" trust models and systems. Such approach is limiting due to variations in technology, conflicts between users' requirements and/or conflicts between user requirements and service outcomes. Additionally, this approach may result in service providers being overwhelmed by adding new resources to satisfy all possible requirements, while having no information or guarantees about the level of trust they gain in the network. Accordingly, we hypothesize the need for personalizable customizable Trust Management Systems (TMSs) for the robustness and wide-scale adoption of large-scale marketplaces for resources and services. Most contemporary TMSs suffer from the following drawbacks: • Oblivious to diversities in trustors' requirements, • Primarily utilize feedback and direct or indirect experience as the only form of credentials and trust computations, • Trust computation methodologies are generally hardcoded and not reconfigurable, • Trust management operations, which we identify as monitoring, data management, analysis, expectation management, and decision making, are tightly coupled. Such coupling impedes customizability and personalization, and • Do not consider context in trust computations, where trust perspectives may vary from a context to another. Given these drawbacks and the large scale of the global marketplace of resources and services, a reference architecture for trust management systems is needed, which can incorporate current systems and may be used in guidance and development of a wide spectrum of trust management systems ranging from un-personalized to fully personalized systems. Up to our knowledge, no TMS reference architecture exists in the literature. In this dissertation, we propose a new Reference Architecture for Trust Management (RATM). The proposed reference architecture applies separation of concern among trust management operations; namely, decision expectation, analytics, data management and monitoring. RATM defines trust management operations through five reconfigurable components which collectively can be used to implement a wide spectrum of trust management systems ranging from generic to highly personalized systems. We used RATM for trust personalization, where we propose a Personalized Trust Management System (PTMS) based on RATM. We evaluated PTMS's scalability and demonstrated its effectiveness, efficiency and resilience by contrasting against a Generic Trust Management System (GTMS). We used two case studies for our evaluations; namely, BitTorrent and a video conferencing application. Intellectual Merit In this work, we propose RATM, a reference architecture for trust management systems that can be used to implement a wide variety of trust management systems ranging from generic systems (un-personalized) to highly personalized systems. We assume service-based environment where consumers, brokers and service providers are interacting and transacting in services and resources to satisfying their own trust requirements. We used RATM to implement a personalized trust management system (TMS). The main contributions of this work are as follows: • Proposing RATM for the guidance and development of a wide spectrum of TMSs ranging from un-personalized to fully personalized systems, and • Utilizing our RATM to propose and implement a personalized, scalable TMS with varying trust computation models. Broader Impact RATM provides reference architecture for trust management which can be used to develop and evaluate a wide spectrum of TMSs. Personalization, in general, paves the road for reaching high levels of satisfaction, where interacting parties' requirements are individually considered and thus consumers are served the best suited service(s). Hence, we claim that PTMS would largely enhance large-scale heterogeneous systems offering services and resources. This could lead to more cooperation, more transactions, more satisfaction, less malicious behavior and lower costs. / PHD

Ethiopia, a developing country on the African continent, has experienced double digit growth during the last decade. To sustain this growth, the country needs a sound and healthy banking sector. Consensus exists that trust is a valuable resource that can make a difference. It influences the quality of relationships between an organisation and its employees and management. It enables organisations such as banks, to retain their most valuable employees and customers, and improve organisational functioning. Most leaders agree that high levels of trust are critical to the success of their organisations. In this study, the existing theoretical principles and models relating to trust, were analysed in depth, and followed up by an empirical study to determine to what extent trust management practices were being applied. A quantitative survey was conducted on a probability sample of 405 participants from the eight largest banks in Ethiopia. From the measured items, 54 usable factors were identified by using exploratory factor analysis. The calculated Cronbach alpha values indicated a satisfactory internal consistency. While the Pearson product-moment correlation applied to the factors, indicated a dominance of statistically significant positive correlations. Various informative results emerged from the empirical survey, amongst which were, the relative weak application of trust management practices such as trust drivers and trust builders. Some barriers were also identified such as the trustworthiness of coworkers/ team members, the immediate supervisor and top management. Thus, it was evident that proper trust management practices had not been fully established within the banking sector in Ethiopia. There was thus a need for some guidelines in this regard. An integrated trust management model was therefore developed, and tested, through Structural Equation Modelling, and validated to satisfy this need. / Human Resource Management / D. Admin. (Human Resource Management)

Distributed systems like e-commerce and e-market places, peer-to-peer networks, social networks, and mobile ad hoc networks require cooperation among the participating entities to guarantee the formation and sustained existence of network services. The reliability of interactions among anonymous entities is a significant issue in such environments. The distributed entities establish connections to interact with others, which may include selfish and misbehaving entities and result in bad experiences. Therefore, trustworthiness evaluation using trust management techniques has become a significant issue in securing these environments to allow entities decide on the reliability and trustworthiness of other entities, besides it helps coping with defection problems and stimulating entities to cooperate. Recent models on evaluating trustworthiness in distributed systems have heavily focused on assessing trustworthiness of entities and isolate misbehaviours based on single trust metrics. Less effort has been put on the investigation of the subjective nature and differences in the way trustworthiness is perceived to produce a composite multidimensional trust metrics to overcome the limitation of considering single trust metric. In the light of this context, this thesis concerns the evaluation of entities’ trustworthiness by the design and investigation of trust metrics that are computed using multiple properties of trust and considering environment. Based on the concept of probabilistic theory of trust management technique, this thesis models trust systems and designs cooperation techniques to evaluate trustworthiness in mobile ad hoc networks (MANETs). A recommendation based trust model with multi-parameters filtering algorithm, and multidimensional metric based on social and QoS trust model are proposed to secure MANETs. Effectiveness of each of these models in evaluating trustworthiness and discovering misbehaving nodes prior to interactions, as well as their influence on the network performance has been investigated. The results of investigating both the trustworthiness evaluation and the network performance are promising.