While collaborative models that integrate human doctors and AI agents are increasingly critical in online healthcare consultations, little is known about how variations in doctors’ professional titles (e.g., senior vs. junior) within these teams influence healthcare consumers’ resistance to AI agents. Drawing on trust transfer theory, this study examines how various compositions of human doctor-AI teams affect consumers’ resistance to the AI agent. Across four scenario-based studies, we discovered that human doctor-AI teaming can reduce consumers’ resistance to the AI agent. Specifically, senior doctor-AI teaming more effectively reduces consumers’ resistance to the AI agent by enhancing trust transfer compared to junior doctor-AI teaming. However, this effect was not observed when the severity of the diseases consulted by consumers was low or the AI agent served as a partner. This study enriches human-AI collaboration research and suggests strategies to improve consumer acceptance of AI in healthcare services.

Video-conferencing (VC)-enabled online learning has become a norm in schools and institutions. However, the multitasking capabilities of VC often lead users to exhibit cyberloafing behaviors in learning contexts. Despite its prevalence, there is limited understanding of such deviant behaviors in VC-enabled online learning and how to address them. To tackle this gap, we draw on conservation of resources (COR) theory to investigate the evolutionary logic of effort investment behaviors. Using a sequential mixed-methods approach, we first conducted semi-structured interviews, which revealed two constraining resources and one empowering resource of VC in learning contexts. We propose that an individual's consumption of such environmental resources, along with their individual resources, leads to gains in and losses of self-regulated learning (SRL) within VC-enabled online learning, which in turn affect subsequent cyberloafing behaviors. To test our research model and hypotheses, we conducted a survey with 309 participants, followed by a post-survey interview to triangulate the findings. Our study offers valuable insights into the design and implementation of VC in learning contexts to mitigate cyberloafing behaviors, contributing to both theory and practice.

The transformation of AI into a collaborative teammate necessitates a shift in focus from task performance to sustainable, human-centered outcomes. A central challenge to this vision is AI misinformation. This study investigates how AI misinformation shapes human contribution by elucidating the mediating role of team situation models (TSMs). We used a mixed-methods approach, beginning with 30 semi-structured interviews to identify context-specific TSMs influenced by AI factual errors. We then proposed a research model and tested it in a vignette-based experiment with 401 participants. We empirically tested the hypotheses using structural equation modeling and multi-group analysis. The results uncover a multi-level framework comprising four key TSMs: teammate-level (trust in the AI teammate and trust in the human teammate) and team-level (collective efficacy and group goal commitment). We find that AI factual errors erode trust at the teammate level, which in turn negatively impacts team-level TSMs. These weakened team-level TSMs subsequently increase individuals’ tendency to withhold effort in human-AI team collaboration. Moreover, when the human teammate (rather than the focal individual) detects the AI misinformation first, this detection buffers AI misinformation’s negative effects on trust in that teammate, strengthens the positive relationship between trust in the AI and collective efficacy, and weakens the link between trust in the human teammate and collective efficacy. Our research yields valuable theoretical and practical implications for stimulating human contribution in human-AI collaboration.

This research explores the implications for risk of six AI focused Fintech business models identified in the previous research. While each of the six business models has been validated in previous research their typical influence on risk has not been sufficiently explored. A Delphi method with two stages is used. In the first stage, the semi-structured interviews with experts identify the main risks each of the six models faces. For the second stage the participants are presented with a summary of all the risks identified by all the participants, and they are asked to rank them for their specific model. Eight risks related to the application of an AI focused business model are identified: (1) direct AI risks, (2) technology integration risk, (3) third-party vendor risk, (4) technology infrastructure risk, (5) data protection and privacy risk, (6) cybersecurity risk, (7) fraud and identity theft risk, and (8) strategic misalignment risk. The eight risks apply to all six models however they are not equally significant in all models. The six business models reduce risk in two main ways: Firstly, because they are proven models and secondly because they have a proven place in the typical ecosystems that emerge.

In the Metaverse era, immersive technologies such as virtual reality (VR) are reshaping collaborations through engaging but complex collaborative virtual environments (CVEs). For this reason, it is essential to identify the key attributes that encourage individuals to use CVEs in such alliances. Grounded in situated cognition theory, this study investigates the key attributes of CVEs and uses trust as a critical mediator to determine how they jointly influence usage intention. First, we identified potential CVE features through a literature review, after which we refined them via a qualitative study to identify those most relevant to usage intention. Building on these insights, we developed a context-specific model that we tested through a sequential quantitative study. Our results reveal a significant interaction between spatial openness and avatar appearance formality in shaping collaboration intention. Importantly, results show that enclosed spaces, especially when paired with formal avatar appearances, enhance perceptions of ability and integrity trust, which in turn positively affect collaboration intention. By clarifying how contextual features shape trust in virtual settings, this study not only explains the underlying mechanisms but also provides practical guidance for designing CVEs through deliberate spatial-avatar configurations.

Increasing concerns about greenhouse gas emissions and local air pollution have motivated the search for sustainable transportation alternatives. Electric vehicles (EVs) offer a promising solution, yet their potential to replace internal combustion engine vehicles (ICEVs) in high-utilization ridesharing fleets remains insufficiently understood. This study analyzes data from a major ridesharing platform in Beijing, comparing EVs and ICEVs across car-sharing and private car-hailing modes. We find that at the driver level, EVs and ICEVs demonstrate comparable order completion in car-sharing, while EVs exhibit significantly lower daily mileage. Order-level regression analysis reveals that EVs, in private ride-sharing, exhibit a higher propensity to complete long-distance orders. We also find that mainstream 2015 EVs could cover the majority of daily ICEV travel patterns, with higher substitutability in car-sharing. These findings provide a benchmark for evaluating progress in the ongoing electrification of shared mobility.

The Metaverse shows significant potential to support collaboration by enabling immersive virtual environments for collective work and learning. However, its effectiveness is often limited by a lack of user trust and engagement . By contextualizing and extending the interactionist theory of place attachment, this study theorizes spatial familiarity as a core feature of Metaverse, examining its mechanisms for fostering perceptions of virtual environments as reliable and engaging places for collaboration. With this end in view, this research followed a sequential mixed-methods approach, with the qualitative method first developing the context-specific research model by identifying key factors that enable Metaverse-based collaboration, and then the quantitative method validating the model. The results together indicate that the spatial familiarity of Metaverse enables individuals to engage in virtual environments for collaboration both directly and mediated by trust. Importantly, the efficacy of spatial familiarity is particularly pronounced when a person perceives a high level of avatar human-likeness but diminishes when her self-regulation is high. By uncovering the theoretical mechanisms through which spatial stimuli contribute to trust-building in collaborative Metaverse contexts, this study also offers practical implications for designing and managing virtual environments for collaborative purposes.

PurposeThe emergence of live streaming commerce has injected promising impetus into rural development and attracted many rural streamers. This study aims to explore the influencing factors of rural streamers’ engagement intentions to help promote the sustainable development of rural live streaming commerce.Design/methodology/approachGrounded in the extended valence framework, this research employs a mixed-methods approach encompassing both qualitative and quantitative methodologies. In the qualitative phase, the authors conduct in-depth interviews with 15 rural streamers, employing data coding techniques to uncover underlying factors. Subsequently, in the quantitative phase, the authors analyze survey data from 282 rural streamers, subjecting hypotheses to validation through structural equation modeling.FindingsThe findings derived from the analysis of both interviews and questionnaires reveal that several platform qualities, including platform rural-aiding support, perceived effectiveness of dispute resolution, perceived interactivity and platform reputation, have a positive effect on trust in the platform and validate the extended valence framework in understanding rural streamers’ live streaming intention. In addition, ties with customers have a moderating effect. Specifically, the stronger the ties with customers, the stronger the positive effect of perceived benefits and the weaker the positive effect of trust in the platform on live streaming intention will be.Originality/valueThis study contributes to the rural live streaming commerce literature and trust research from the sellers’ perspective and provides practical implications for policymakers and live streaming platform managers on enhancing rural streamers’ participation.

The advent of generative artificial intelligence (GAI) increasingly transforms artificial intelligence (AI) models from passive tools to proactive social actors. Organizations are increasingly deploying GAI as a teammate to improve work efficiency and decrease costs. However, this promising and sustainable development of GAI is uncertain because its output possesses much misinformation. Inadequate reliability and accuracy of GAI misinformation may distrust of human teammates and elicit unsatisfied collaboration outcomes. Decreasing over-reliance and over-trust may also offer the opportunity to dilute the harm of AI on human unique knowledge and activate innovative thoughts. Given these mixed theorizations, this paper leverages affordance actualization theory to draw a comprehensive understanding of misinformation in human-GAI team collaboration. Leveraging a mixed-methods design, we aim to begin with a qualitative study to identify the affordances of AI and interpret the actualization process. Based on the results, we intend to conduct a Wizard of Oz experiment to examine the proposed model.

PurposeDue to ChatGPT's relative novelty, the determinants and consequences of trust in it remain largely unexplored, especially in the field of higher education. This research investigates how perceptions of AI can foster university students' trust in ChatGPT, which in turn leads to their continuance and sharing intentions. Cross-cultural differences are also analyzed to provide culturally relevant insights.Design/methodology/approachA mixed-methods approach was employed. In the qualitative phase (Study 1), we interviewed 30 university students and used an inductive methodology to analyze the data. In the follow-up quantitative studies, the hypotheses were examined using a sample of 249 students from China (Study 2) and 204 students from the United States (Study 3).FindingsQualitative results suggest that perceived benefits (convenience, usefulness and controllability) and perceived risks (inflexibility, privacy invasion and overreliance) are identified as antecedents of learners' trust in ChatGPT. Quantitative results reveal that factors including convenience, usefulness and overreliance on AI positively influence learners' trust in ChatGPT. Furthermore, trust positively affects learners' continuance intention and sharing intention. These findings are consistent across the Chinese and American samples. However, inflexibility significantly influences trust among Chinese learners but not among American learners. Additionally, invasion of privacy negatively affects trust in ChatGPT only for the US sample, suggesting cultural divergence in trust formation between these two countries.Originality/valueWe shed light on the antecedents and consequences of trust in ChatGPT within higher education settings, especially across different cultures. This research endeavor also promotes the integration of ChatGPT for learning activities to maximize the utility of this technology.

Healthcare question-answering (QA) systems can assist physicians in making medical decisions. However, traditional medical QA systems struggle with multi-agents interaction and domain-specific knowledge processing, thereby reducing the accuracy and credibility of clinical decision-making. We thus develop a multi-agent decision-making system by combining a fine-tuned medical model, biomedical knowledge graphs, and PubMed data. By summarizing the symptoms described by users, our system can automatically convene clinical experts from various fields, retrieve domain knowledge, and provide clinical decision support for users. We have validated the system performance using both technical and user-centric approaches in terms of information accuracy, user satisfaction, user trust, ect. We thus provide an effective tool for healthcare professionals to make accurate and timely decisions. Furthermore, this study also reveals new design and research opportunities, including (1) optimizing multi-agent collaboration mechanisms for more complex medical decision-making, (2) improving interaction design to enhance system transparency and explainability, and (3) expanding the system to support a broader range of medical issues and multimodal data.

The emergence of the sharing economy has catalyzed the proliferation of online knowledge payment platforms, empowering individuals possessing untapped knowledge to engage as sellers. Within this landscape, reputation plays a pivotal role in shaping the sales dynamics of knowledge products, which are categorized as credence goods. Unlike conventional sellers, who typically focus on internal reputation-building efforts, knowledge providers have the opportunity to cultivate an external reputation through their professional contributions beyond the platform, potentially influencing sales within it. However, the relative impact of internal and external reputation on driving sales remains uncertain. Our study addresses this gap by investigating the influence of a knowledge provider's internal and external reputation on sales across two phases of live knowledge-sharing lectures—pre and post-live broadcast—using data obtained from a prominent online knowledge payment platform. Our findings reveal that both internal and external reputation positively impact sales during both phases. Furthermore, our analysis indicates that the type of product moderates the effect of reputation on sales, with internal reputation exerting a greater influence on sales of hedonic products, while external reputation has a stronger effect on sales of utilitarian products. These discoveries contribute to technology operations management by providing theoretical insights into the formulation and implementation of innovative strategies aimed at enhancing the sales performance of knowledge-based products. Our findings underscore the importance of integrating distinct reputation-building approaches into the design of reputation systems, thereby charting a course toward achieving successful product sales in the realm of online knowledge payment platforms.

Communication and reflection abilities are critical in managing strategic cooperation between humans and Generative Artificial Intelligence (GAI), especially when facing conflict in decision-making. This study introduces two variations of EmotionPrompt strategies, drawing on regulatory focus theory, to explore both individuals' perceptions of GAI ability and their empowerment in self-competence when handling disagreements. An experiment between humans and GAI chatbots in determining product promotion strategy showed that emotional prompts impact individuals' reappraisals of both chatbots and their own performance profoundly, cultivating self-determination in the final decision. Importantly, EmotionPrompt with promotion orientation can increase the perceived flexibility of chatbot decision-makers, facilitating individual self-enhancement and trust in GAI competence. In contrast, the prevention-oriented EmotionPrompt appears to constrain individuals' judgments and decision-making processes, as evidenced by the increased occurrence of inhibit words and anxiety emotions in their reflections. These findings provide novel perspectives on implementing specific regulatory-oriented EmotionPrompt strategies in GAI to address opinion conflicts in decision-making with humans.

Central bank digital currencies (CBDC) have been implemented by some countries and trialled by many more. The consumer has an increasing range of financial services to choose from including decentralised blockchain-based cryptocurrencies. A CBDC may use blockchain technology, but it is centralized, so the institutions that support it play an important role. Despite the centralised top-down nature of this financial technology, it still needs to be adopted so the consumer’s perspective, particularly their trust in it, is very important. Each CBDC implementation can be different, and each country’s context can be different, therefore it is important to understand each case separately. This research models the Brazilian consumer’s trust in their two-tier CBDC, where the central bank and the retail banks retain their current role. The six ways to build trust in a CBDC, identified by previous research in a different region, are supported for this case also. These are: (a) Trust in government and central bank offering the CBDC, (b) expressed guarantees for those using it, (c) the favourable reputation of other active CBDCs, (d) the CBDC technology, the automation and limited human involvement necessary, (e) the trust building features of the retail bank’s CBDC wallet app, and (f) the privacy features of the retail bank’s CBDC wallet app and back-end processes.

Companies in various industries are attempting to integrate Generative Artificial Intelligence (GAI) into their existing businesses. In the e-commerce domain, GAI has shown tremendous potential in generating online reviews. However, existing literature has paid less attention to how consumers respond to GAI-generated reviews versus human-generated reviews. Moreover, little research has explored whether and why consumers are willing to use GAI to generate online reviews. By conducting two experiments, this study investigates how consumers respond differently to GAI-generated reviews versus human-generated reviews and identifies potential factors that influence consumers’ willingness to use GAI to generate reviews. Findings indicate that although there is no significant difference in consumers’ perceptions between human-generated and GAI-generated reviews in terms of review credibility, review richness, and review usefulness, only half of the participants are willing to use GAI to generate reviews. Further analysis results suggest that individuals who consider GAI unethical tend to avoid using GAI. Those with high personal innovativeness are more willing to use GAI to generate online reviews. Our findings deepen the understanding of consumer attitudes toward GAI-generated reviews and provide implications for the deployment of GAI in the online review system.

The presence of an AI teammate changes traditional collaboration dynamics found in human-human teams. Trust is especially important for human members of human-AI teams (HATs) to cope with changes in collaborative environments. However, little research focuses on trust development in HATs. The process of trust development can be viewed as a process of initial trust expectations being met or unmet. Drawing upon expectation disconfirmation theory, this study investigates how trust expectation disconfirmation influences trusting intention and willingness to work with the focal teammate (AI or human). Results indicate that as positive disconfirmation increases or negative disconfirmation decreases, individuals' trusting intention toward the human teammate will increase. However, both negative and positive disconfirmation is harmful to trusting intention toward the AI teammate. We also confirm the mediating effect of trusting intention on the impact of expectation disconfirmation on willingness to work with the focal teammate.

The development of Artificial General Intelligence (AGI) in e-commerce has garnered significant attention. AGI's cross-domain cognitive capabilities are revolutionizing e-commerce through personalized shopping experiences, intelligent customer service, and efficient supply chain management. However, challenges like algorithm aversion, ethical concerns, and data privacy issues persist. This special issue explores AGI's effects and challenges in ecommerce, aiming to advance understanding of its potential and address problems hindering adoption. Contributions cover topics such as AGI's impact on consumer behavior, sales performance, and employee performance, as well as strategies for responsible AGI deployment.