This is a bilingual snapshot page saved by the user at 2024-6-11 18:26 for https://app.immersivetranslate.com/pdf-pro/4ffdb559-f786-4652-a61f-29084d303850, provided with bilingual support by Immersive Translate. Learn how to save?
2024_06_11_a3e5dc861b8dbbd23eb3g

1.引言  1. Introduction

工业 5.0 的兴起推动了物流行业作为其他产业重要服务领域的蓬勃发展 [1]。然而随着城市化进程的加速和人口的不断增长, 交通运输压力正在逐渐增大,传统的交通运输方式已难以满足日益增长的需求。此外, 运输需求的增加、运输成本的不断上升、供应链复杂性、交通安全风险等问题也愈发凸显[2-4]。因此, 物流行业迫切需要采取一种先进的数字化战略, 以应对众多棘手挑战。特别是考虑到可持续发展基础设施需求不断增加[5], 数字化转型已成为现代物流行业发展的必然趋势[1]。
The rise of Industry 5.0 has driven the flourishing development of the logistics industry as an important service sector for other industries. However, with the acceleration of urbanization and continuous population growth, the pressure on transportation is gradually increasing, and traditional transportation methods are struggling to meet the growing demands. In addition, issues such as increasing transportation needs, rising transportation costs, supply chain complexity, and traffic safety risks are becoming more prominent. Therefore, the logistics industry urgently needs to adopt an advanced digital strategy to address numerous challenging issues. Especially considering the increasing demand for sustainable development infrastructure, digital transformation has become an inevitable trend in the modern logistics industry.
在此背景下,人工智能(AI)作为一项引领性的战略技术,正在引领新一轮技术革命和产业升级的浪潮。在推动传统物流行业实现转型的进程中, AI 展现出了至关重要的作用[6]。与传统方法相比, 人工智能技术的采用显著提升了自动化水平, 其解决问题的能力在准确性、速度和输入量上均得到大幅提升 [1],进而显著增强了企业的竞争优势。相关文献证明,在全球气候变暖的严峻挑战下, 物流行业作为碳排放的主要来源之一, 其绿色转型显得尤为重要[1]。在这一转型过程中, 人工智能的应用成为了关键驱动力, 为物流行业的绿色经济发展铺平了道路[7]。在现有学术研究中, 供应链管理已被广泛视为最具潜力从人工智能技术中汲取显著益处的领域之一[8]。通过深入剖析相关文献,我们观察到人工智能技术不仅能为物流运营带来了显著的效率提升, 更在创造社会价值方面展现出巨大潜力[9]。人工智能技术为物流行业提供了以贡献为导向的广阔机遇[10], 预示着未来供应链管理的革新与发展。
Against this background, artificial intelligence (AI) as a leading strategic technology is leading a new wave of technological revolution and industrial upgrading. In the process of promoting the transformation of the traditional logistics industry, AI has shown a crucial role. Compared to traditional methods, the adoption of artificial intelligence technology has significantly improved the level of automation, with its problem-solving capabilities greatly enhanced in terms of accuracy, speed, and input volume, thereby significantly strengthening the competitive advantage of enterprises. Relevant literature proves that in the face of the severe challenge of global warming, the logistics industry, as one of the main sources of carbon emissions, the green transformation is particularly important. In this transformation process, the application of artificial intelligence has become a key driving force, paving the way for the green economic development of the logistics industry. In existing academic research, supply chain management has been widely regarded as one of the most promising areas to derive significant benefits from artificial intelligence technology. Through in-depth analysis of relevant literature, we observe that artificial intelligence technology not only brings significant efficiency improvements to logistics operations but also demonstrates great potential in creating social value. Artificial intelligence technology provides the logistics industry with broad opportunities oriented towards contribution, heralding the innovation and development of future supply chain management.
尽管人工智能技术在物流行业中具有巨大的应用潜力, 但其在该行业内的实际开发和应用程度尚未达到预期水平[11], 这主要归因于行业内多数从业者尚未充分准备采纳这一前沿技术[12]。在过去的几十年里, 人工智能技术在自动化驾驶领域已经进行了许多开发, 具体体现在自动驾驶汽车、智能驾驶公共交通的应用和无人驾驶航空运输等多个方面。同时, 在智能物流系统与供应链可持续化管理系统中,人工智能技术也发挥着不可或缺的作用 。然而,尽管人工智能技术在物流行业已经取得了上述进展, 但其开发程度仍然仅处于初级阶段, 并且存在许多问题和质疑。例如自动驾驶技术面临安全性验证及商业化难题, 货物追踪系统智能化不足, 智能交通管理系统整合有限等问题 ,
Although artificial intelligence technology has great potential for application in the logistics industry, its actual development and application in the industry have not yet reached the expected level. This is mainly due to the fact that most practitioners in the industry are not fully prepared to adopt this cutting-edge technology. Over the past few decades, artificial intelligence technology has made many developments in the field of automated driving, specifically in the applications of autonomous vehicles, intelligent public transportation, and unmanned aerial transportation, among others. At the same time, artificial intelligence technology also plays an indispensable role in intelligent logistics systems and sustainable supply chain management systems. However, despite the progress made by artificial intelligence technology in the logistics industry, its development is still in the early stages and there are many issues and doubts. For example, autonomous driving technology faces challenges in safety verification and commercialization, insufficient intelligence in cargo tracking systems, and limited integration in intelligent traffic management systems.
15]。由此可见, 人工智能在物流行业的应用趋势、未来机遇和应用障碍方面存在灰色地带。
15]. It can be seen that there are gray areas in the application trends, future opportunities, and application barriers of artificial intelligence in the logistics industry.
自 2020 年起, 全球每年对行业数字化的投资预计将达到 9000 亿美元。作为全球物流行业最多的国家之一, 中国在这一领域的表现尤为突出。据 2023 年数据显示,中国的物流行业作为服务业的关键组成部分,其总产值已达到 13.2 万亿元人民币,同比增长 。虽然中国物流收入规模总体呈现持续扩张的趋势, 但物流成本的占比仍然较高。根据相关文献,中国物流成本占 GDP 的比重仍然较高, 与国际先进水平相比还有一定差距[16]。这其中,信息不对称、流程繁琐等问题是制约物流成本降低的重要因素。近年来,中国公路货运的单车运营效率持续下滑, 空驶率、车辆维修率等成本不断上升[17]。同时, 水路运输的智能化水平也亟待提升, 船舶运营效率、航线规划等方面存在诸多优化空间[18]。这与中国物流行业内部数字化水平不高、信息透明度不足密切相关[11]。因此, 面对这些挑战和机遇, 中国物流行业必须加快数字化转型的步伐。通过引入先进的信息技术、大数据、人工智能等数字化手段, 提高运输效率、降低运营成本、优化资源配置[9], 从而应对日益激烈的市场竞争。同时, 数字化转型还可以通过提升服务质量、增强客户体验、提高透明度, 来为行业赢得更多市场份额[19]。
Since 2020, global annual investment in industry digitization is expected to reach $900 billion. As one of the countries with the largest logistics industry in the world, China's performance in this field is particularly outstanding. According to data from 2023, China's logistics industry, as a key component of the service industry, has reached a total output value of 13.2 trillion yuan, a year-on-year increase of . Although the overall scale of logistics revenue in China continues to expand, the proportion of logistics costs remains high. According to relevant literature, the proportion of logistics costs to GDP in China is still relatively high, with a certain gap compared to international advanced levels. Among them, issues such as information asymmetry and cumbersome processes are important factors restricting the reduction of logistics costs. In recent years, the operational efficiency of road freight transportation in China has continued to decline, with costs such as empty running rate and vehicle maintenance rate constantly rising. At the same time, the level of intelligence in waterway transportation also needs to be improved urgently, with many optimization opportunities in ship operation efficiency and route planning. This is closely related to the low level of digitization and insufficient information transparency in the Chinese logistics industry. Therefore, faced with these challenges and opportunities, the Chinese logistics industry must accelerate the pace of digital transformation. By introducing advanced information technology, big data, artificial intelligence, and other digital means, improving transportation efficiency, reducing operating costs, optimizing resource allocation, and coping with increasingly fierce market competition. At the same time, digital transformation can also win more market share for the industry by improving service quality, enhancing customer experience, and increasing transparency.
与以往的研究不同, 本研究重点关注微观层面, 研究采用的模型融合了 UTAUT 模型中绩效期望、努力期望、社会影响和便利条件四个核心维度[20],并额外考虑了人工智能焦虑和个人创新性两个个体层面因素。在深入探讨物流行业中人工智能技术的采纳行为时, 虽然既有的研究已将 UTAUT 及其扩展模型引入该领域, 但多数研究主要聚焦于组织层级的采纳动态 。然而, 我们必须强调, 员工个体层级的采纳是组织整体采纳人工智能技术不可或缺且前置的关键环节。与组织采纳的宏观视角不同, 员工采纳过程中可能涉及一系列独特的因素, 这些因素在塑造员工对人工智能技术的态度、感知和采纳行为上发挥着至关重要的作用, 特别是在识别和解决采纳障碍方面[22]。所以, 本研究致力于通过应用一个经过精心拓展的 UTAUT 模型, 以更为严谨和系统的学术视角, 全面解析物流行业中用户采纳人工智能技术的复杂机制与动态。因此,本研究旨在实现以下目标:
Unlike previous studies, this study focuses on the micro-level and integrates the UTAUT model's four core dimensions of performance expectancy, effort expectancy, social influence, and facilitating conditions, as well as considering individual factors such as artificial intelligence anxiety and personal innovativeness. While existing research has introduced the UTAUT and its extended models into the field of logistics industry's adoption behavior of artificial intelligence technology, most studies have mainly focused on organizational-level adoption dynamics. However, it is essential to emphasize that individual-level adoption by employees is a crucial and prerequisite step for the overall organizational adoption of artificial intelligence technology. Different from the macro perspective of organizational adoption, the employee adoption process may involve a series of unique factors that play a vital role in shaping employees' attitudes, perceptions, and adoption behaviors towards artificial intelligence technology, especially in identifying and addressing adoption barriers. Therefore, this study aims to comprehensively analyze the complex mechanisms and dynamics of user adoption of artificial intelligence technology in the logistics industry from a more rigorous and systematic academic perspective by applying an elaborated UTAUT model. Thus, this study aims to achieve the following objectives:
[1] 在扩展 UTAUT 框架的基础上, 确定物流行业采用人工智能技术的主要影响因素。
[1] On the basis of extending the UTAUT framework, determine the main influencing factors of the logistics industry adopting artificial intelligence technology.
[2] 确定如何推动物流行业的个人用户采用人工智能技术。
[2] Determine how to promote individual users in the logistics industry to adopt artificial intelligence technology.
本文的其余部分结构如下:第 2 节概述了研究模型并发展了研究假设。第 3 节介绍了研究方法, 包括问卷设计、数据收集与数据统计。第 4 节对数据进行了分析并得到了结果, 之后在第 5 节进行了讨论。最后在第 6 节对本文的研究成果进行总结。
The rest of this article is structured as follows: Section 2 outlines the research model and develops research hypotheses. Section 3 introduces the research methods, including questionnaire design, data collection, and data analysis. Section 4 analyzes the data and presents the results, followed by a discussion in Section 5. Finally, Section 6 summarizes the research findings.

2.理论模型和研究假设 2. Theoretical Model and Research Hypotheses

2.1 理论模型 2.1 Theoretical Model

UTAUT 模型(Unified Theory of Acceptance and Use of Technology)是一种整合性的科技接受模型, 旨在全面解析和预测个体对新技术采纳和使用的影响因素[23]。该模型由 Venkatesh 等人于 2003 年提出[20], 基于多个经典技术接受理论的核心要素, 如理性行为理论(TRA)、技术接受模型(TAM)等, 形成了一个综合性的理论框架。UTAUT 模型主要包括四个核心构造:期望性、努力性、社会影响和设备特性, 这些构造可以进一步细化为绩效期望、努力期望、社会影响和便利条件等因素。UTAUT 模型对技术采纳和使用行为的解释能力高达 70%, 使得 UTAUT 能够在技术采纳和使用的研究上被有效应用[24]。此外, UTAUT 模型还考虑了一些调节因素,如性别、年龄、经验等,这些因素可以影响个体对技术采用的态度和行为。
The UTAUT model (Unified Theory of Acceptance and Use of Technology) is an integrative model of technology acceptance, aimed at comprehensively analyzing and predicting the factors influencing individuals' adoption and use of new technologies. The model was proposed by Venkatesh et al. in 2003, based on core elements of several classic technology acceptance theories such as Theory of Reasoned Action (TRA) and Technology Acceptance Model (TAM), forming a comprehensive theoretical framework. The UTAUT model mainly consists of four core constructs: performance expectancy, effort expectancy, social influence, and facilitating conditions, which can be further refined into factors such as performance expectations, effort expectations, social influence, and facilitating conditions. The UTAUT model has an explanatory power of up to 70% for technology adoption and usage behavior, making it effectively applicable in research on technology adoption and usage. In addition, the UTAUT model also considers some moderating factors such as gender, age, experience, which can influence individuals' attitudes and behaviors towards technology adoption.
在物流行业, 人工智能焦虑指的是由于人工智能技术的引入和应用, 对物流行业从业者、管理者以及相关利益方在整体情感上产生的焦虑或恐惧[25]。在探讨人工智能技术在物流行业应用开发程度低的原因时, 研究发现失业担忧、技术成本、公众信任缺失、安全顾虑、隐私与伦理问题以及技术限制是主要障碍。这些障碍实质上可归因于人工智能焦虑的不同方面, 包括工作替代焦虑、社会技术疑虑、AI 配置困扰以及人工智能学习焦虑[26]。鉴此, 我们将人工智能焦虑作为一个因素纳入模型中加以考量。另一方面, 个人创新性指的是在特定环境中,通过独特的思考和实践,产生新颖、有价值想法或成果的能力[27]。研究个人创新性与新技术采纳之间的关系, 可以更深入地理解物流行业中用户对人工智能技术采用的态度和行为[28]。这种扩展模式有助于区分不同用户群
In the logistics industry, artificial intelligence anxiety refers to the anxiety or fear experienced by practitioners, managers, and related stakeholders in the logistics industry as a whole due to the introduction and application of artificial intelligence technology. When discussing the reasons for the low level of application development of artificial intelligence technology in the logistics industry, studies have found that unemployment concerns, technology costs, lack of public trust, security concerns, privacy and ethical issues, and technological limitations are the main obstacles. These obstacles can essentially be attributed to different aspects of artificial intelligence anxiety, including job replacement anxiety, social technical doubts, AI deployment troubles, and artificial intelligence learning anxiety. Therefore, we incorporate artificial intelligence anxiety as a factor into the model for consideration. On the other hand, personal innovativeness refers to the ability to generate novel and valuable ideas or outcomes through unique thinking and practice in a specific environment. Studying the relationship between personal innovativeness and the adoption of new technologies can provide a deeper understanding of users' attitudes and behaviors towards the adoption of artificial intelligence technology in the logistics industry. This expanded model helps differentiate between different user groups.

体, 以满足物流行业内不同的需求和关注点。尽管个人创新性尚未成为技术接受模型的核心构成要素,但其已被广泛认可,并被视为其他组织采纳新产品或创新的重要预测指标[22]。这两个创新维度的融入显著增强了模型的预测准确性,并从微观层面更深入地揭示了工业 5.0 新兴技术创新治理背景下影响物流行业用户接受人工智能的机制。
Body, to meet different needs and focus points within the logistics industry. Although individual innovativeness has not yet become a core constituent element of the technology acceptance model, it has been widely recognized and considered as an important predictor for other organizations to adopt new products or innovations. The integration of these two dimensions of innovation significantly enhances the predictive accuracy of the model and provides a deeper insight into the mechanisms influencing the acceptance of artificial intelligence by logistics industry users at the micro level under the background of industrial 5.0 emerging technology innovation governance.

2.2 研究假设 2.2 Research Hypothesis

2.2.1 绩效期望与使用行为 2.2.1 Performance Expectations and Usage Behavior

绩效期望被 Venkatesh 定义为个人相信使用某项技术或系统将帮助他获得工作绩效收益的程度。其他模型中与绩效期望有关的结构包括感知有用性、外在动机、工作适合性、相对优势、和结果预期[20], 过去的文献表明,当涉及到新技术的强制和自愿使用时,绩效期望与技术使用之间存在密切的联系[20, 29, 30],现将其扩展到支持人工智能技术在物流行业中的应用也是合乎逻辑的。在人工智能在物流行业应用的背景下,绩效期望意味着人工智能技术可以为其用户带来的好处,包括运输效率、成本、安全性以及服务质量等方面的改善和提升[14], 这些好处会影响用户对使用人工智能技术的积极程度。Kwak 等人在研究中提到, 参与者的绩效期望与他们使用人工智能健康技术的行为呈正相关 [31]。同时, Lanhui 等人显示,绩效期望通过信任度和行为意图的影响,间接影响了公众对自动驾驶公交车使用行为的接受程度[13]。基于上述分析, 本研究假设:
Performance expectancy is defined by Venkatesh as the extent to which an individual believes that using a particular technology or system will help them gain job performance benefits. Other structures related to performance expectancy in various models include perceived usefulness, external motivation, job fit, relative advantage, and outcome expectations. Past literature indicates a close relationship between performance expectancy and technology use when it comes to both mandatory and voluntary use of new technologies. Extending this to support the application of artificial intelligence technology in the logistics industry is also logical. In the context of artificial intelligence application in the logistics industry, performance expectancy implies the benefits that artificial intelligence technology can bring to its users, including improvements in transportation efficiency, costs, safety, and service quality. These benefits will influence the users' positive attitude towards using artificial intelligence technology. Kwak et al. mentioned in their study that participants' performance expectancy is positively correlated with their use of artificial intelligence health technology. At the same time, Lanhui et al. demonstrated that performance expectancy indirectly influences the public's acceptance of using autonomous buses through its impact on trust and behavioral intent. Based on the above analysis, this study assumes:

H1:绩效期望正向影响物流行业用户对人工智能技术的使用行为。
H1: Performance expectations positively influence logistics industry users' use of artificial intelligence technology.

2.2.2 努力期望和使用行为 2.2.2 Efforts to Expect and Use Behavior

努力期望被 Venkatesh 定义为与系统使用相关的轻松程度。来自现有模型的三个构造捕获了努力期望的概念:感知易用性、复杂性和易用性[20]。本研究
Effort expectancy is defined by Venkatesh as the perceived ease of use related to system usage. Three constructs from existing models capture the concept of effort expectancy: perceived ease of use, complexity, and usability [20]. This study
中,努力期望指物流行业用户主观感知下人工智能技术的易用程度 [32], 当用户认为人工智能技术操作简单,快捷方便,且不需要付出过多努力时,他们就更有可能接受并使用这些技术[33]。Lanhui 等人发现,自动系统的易用性将驾驶对用户对其的使用行为产生积极影响[13]。Ali 等人进行的一项研究对巴基斯坦绿色信息技术的采用情况进行了调查, 发现努力期望对消费者的行为意图有着显著影响[34], Sharafi 等人进一步得出结论: 努力期望对消费者使用绿色产
In the logistics industry, efforts are made to expect the subjective perception of the ease of use of artificial intelligence technology by users [32]. When users perceive that artificial intelligence technology is easy to operate, quick and convenient, and does not require too much effort, they are more likely to accept and use these technologies [33]. Lanhui et al. found that the usability of automated systems will have a positive impact on users' usage behavior [13]. A study conducted by Ali et al. investigated the adoption of green information technology in Pakistan and found that effort expectancy has a significant impact on consumer behavior intentions [34]. Sharafi et al. further concluded that effort expectancy affects consumer use of green products.

品的行为产生积极影响[35]。基于上述分析, 本研究假设:
The behavior of the product has a positive impact. Based on the above analysis, this study assumes:

H2:努力期望正向影响物流行业用户对人工智能技术的使用行为。
H2: Strive to positively influence logistics industry users' use of artificial intelligence technology.

2.2.3 努力期望和绩效期望 2.2.3 Effort Expectation and Performance Expectation

在 UTAUT 模型中,努力期望和绩效期望是相互关联的[20]。研究显示,新技术的高易用性会提升用户对其深入了解和探索的意愿,进而发现其潜在的绩效提升价值[31]。当用户认为使用人工智能技术相对容易时, 他们就会认为引入人工智能可以自动化处理繁琐、复杂的任务, 从而提高工作效率。Patil 的研究发现, 智能手机的使用门槛越低, 老年人对其就有越好的印象, 认为其效能越高[36]。同时, Norzelan 的研究也表明, 如果共享服务行业财务和会计部门人员发现人工智能技术易于使用, 他们就更有可能采用并将其融入日常工作中,这可能会导致该技术的使用量增加和价值更大[37]。基于上述分析,本研究假设:
In the UTAUT model, effort expectancy and performance expectancy are interrelated. Research shows that high usability of new technologies will increase users' willingness to explore and understand them in depth, thereby discovering their potential performance improvement value. When users perceive the use of artificial intelligence technology as relatively easy, they believe that introducing artificial intelligence can automate tedious and complex tasks, thereby improving work efficiency. Patil's study found that the lower the threshold for using smartphones, the better impression elderly people have of them, believing that their effectiveness is higher. At the same time, Norzelan's study also indicates that if financial and accounting personnel in the shared services industry find artificial intelligence technology easy to use, they are more likely to adopt and integrate it into their daily work, which may lead to an increase in the use of this technology and greater value. Based on the above analysis, this study assumes:

H3: 努力预期望正向影响用户对物流行业人工智能技术的绩效期望。
H3: Efforts are expected to positively influence users' performance expectations of artificial intelligence technology in the logistics industry.

2.2.4 社会影响和使用行为 2.2.4 Social Impact and Usage Behavior

社会影响被 Venkatesh 定义为个体感知到的社会压力对其是否采用新技术的影响。与社会影响相关的三个结构为:主观规范、社会因素和形象[20]。个体在决策是否采纳一项新技术时, 往往会受到其社会网络(如意见领袖、同事、家人等)对该技术的看法和态度的影响,当个体感受其积极态度时,他们就更有可能受到这种正面社会影响,从而产生积极的使用行为[38, 39]。由于人工智能产品在物流行业属于新兴产品[40], 用户可能缺乏判断使用人工智能产品是否合适的知识或能力。因此,他们可能会根据其社会群体的态度来形成对使用人工智能产品的意愿和行为[41]。例如, Dogan Gursoy 等指出, 当酒店顾客发现周围大多数人都在使用人工智能产品时, 他们也将更有可能使用这些产品 [42]。基于上述分析,本研究假设:
Social impact is defined by Venkatesh as the influence of perceived social pressure on individuals' adoption of new technologies. Three structures related to social impact are subjective norms, social factors, and image. When individuals decide whether to adopt a new technology, they are often influenced by the opinions and attitudes of their social network (such as opinion leaders, colleagues, family, etc.) towards that technology. When individuals perceive a positive attitude, they are more likely to experience this positive social impact and thus engage in positive usage behavior. As artificial intelligence products are considered emerging products in the logistics industry, users may lack the knowledge or ability to judge whether using artificial intelligence products is appropriate. Therefore, they may form their willingness and behavior to use artificial intelligence products based on the attitudes of their social groups. For example, Dogan Gursoy and others pointed out that when hotel customers see that most people around them are using artificial intelligence products, they are more likely to use these products as well. Based on the above analysis, this study assumes:

H4:社会影响力正向影响物流行业用户对人工智能技术的使用行为。
H4: Social influence positively affects logistics industry users' use of artificial intelligence technology.

2.2.5 便利条件和使用行为 2.2.5 Convenience conditions and usage behavior

便利条件被 Venkatesh 定义为个人相信存在支持系统使用的组织和技术基础设施的程度。这个定义捕捉了三种不同结构所体现的概念:感知行为控制、促进条件和兼容性[20]。在本研究中,当物流行业的用户认为他们有足够的资源
Convenience conditions are defined by Venkatesh as the extent to which individuals believe in the existence of organizational and technological infrastructure that supports system use. This definition captures the concepts embodied in three different structures: perceived behavioral control, facilitating conditions, and compatibility. In this study, when users in the logistics industry believe they have sufficient resources.

和支持来使用人工智能技术时,他们就更有可能接受并使用这些技术[13]。过去的文献表明, 便利条件在塑造用户对于特定技术使用行为的积极态度方面发挥着关键性影响[43]。特别是在针对人工智能使用的研究中, 研究者们发现便利条件与用户使用人工智能的行为之间呈现出显著的正向关联性[44]。近几年 Azman Ong 等人在研究中提到, 如果经过适当的培训和实践经验后, 农村居民拥有更为优越的便利条件,那么他们采用数字支付的可能性就会增加[45]。同时, Khogali 等人也证实了人工智能技术能够降低用户的使用门槛和成本, 提升了用户的使用体验与满意度,从而对用户使用人工智能技术的行为有积极的影响[46]。基于上述分析, 本研究假设:
When people are supported to use artificial intelligence technology, they are more likely to accept and use these technologies. Past literature has shown that convenience conditions play a crucial role in shaping users' positive attitudes towards specific technology use behaviors. In particular, in research on the use of artificial intelligence, researchers have found a significant positive correlation between convenience conditions and user behaviors in using artificial intelligence. In recent years, Azman Ong and others mentioned in their research that if rural residents have better convenience conditions after appropriate training and practical experience, their likelihood of adopting digital payments will increase. At the same time, Khogali and others have also confirmed that artificial intelligence technology can reduce users' barriers and costs, improve users' experience and satisfaction, and thus have a positive impact on users' behaviors in using artificial intelligence technology. Based on the above analysis, this study assumes:

H5: 便利条件正向影响物流行业用户对人工智能技术的使用行为。
H5: Convenient conditions positively influence logistics industry users' usage behavior of artificial intelligence technology.

2.2.6 个人创新性和绩效期望 2.2.6 Personal Innovation and Performance Expectations

Cudjoe 和 Wang 认为个人创新性可以被视为个体在采纳和使用新技术时, 所展现出的对创新更为积极的态度[27]。多项研究证实了个人创新性作为外部变量在解释技术接受领域中的感知有用性和感知易用性方面的作用 。通过研究发现, 创新用户通常扮演着新技术的早期采纳者角色, 他们对新技术表现出坚定不移的信任。即便在技术的潜在价值和具体利益尚未明确的情况下, 这些创新用户依然对新技术有很高的期望[48]。Cudjoe 等人发现,个人创新性对感知的有用性、感知的易用性和中国消费者对碳足迹监测应用的绩效期望产生积极影响[27]。本研究现拓展至物流领域,尽管行业内用户对人工智能的价值与利益持不确定性, 但具有高度创新性的用户更容易认识到人工智能对于提高工作效率、增加行业竞争力等方面的潜力。这种感知有用性的提升会增强他们对人工智能技术的期望[37]。基于上述分析, 本研究假设:
Cudjoe and Wang believe that individual innovativeness can be seen as a more positive attitude towards innovation displayed by individuals when adopting and using new technologies. Multiple studies have confirmed the role of individual innovativeness as an external variable in explaining the perceived usefulness and perceived ease of use in the field of technology acceptance. Research has found that innovative users typically play the role of early adopters of new technologies, showing unwavering trust in them. Even in situations where the potential value and specific benefits of the technology are not clear, these innovative users still have high expectations for new technologies. Cudjoe et al. found that individual innovativeness has a positive impact on perceived usefulness, perceived ease of use, and performance expectations of Chinese consumers for carbon footprint monitoring applications. This study now extends to the logistics field, where despite uncertainty among industry users about the value and benefits of artificial intelligence, highly innovative users are more likely to recognize the potential of artificial intelligence in improving work efficiency and increasing industry competitiveness. This enhancement of perceived usefulness will strengthen their expectations for artificial intelligence technology. Based on the above analysis, this study assumes:

H6:个人创新性正向影响用户对物流行业人工智能技术的绩效期望。
H6: Personal innovativeness positively influences users' performance expectations of artificial intelligence technology in the logistics industry.

2.2.7 个人创新性和使用行为 2.2.7 Personal Innovation and Usage Behavior

Kwarteng 等人在研究中提到, 个人文化价值观在通过心理特征影响一门新技术的未来使用方面发挥着极其重大的作用[49], 并且个人创新性与个人文化价值观有着十分密切的联系[50], 人工智能在物流行业也是一项新兴技术, 由此可以推断, 个人创新性对用户使用人工智能技术行为有很大的影响。Patil 等人显示,就消费者采用移动支付的行为而言,消费者的个人创新性将发挥积极
Kwarteng et al. mentioned in their research that personal cultural values play a significant role in influencing the future use of a new technology through psychological characteristics, and there is a very close relationship between personal innovativeness and personal cultural values. Artificial intelligence is also an emerging technology in the logistics industry, so it can be inferred that personal innovativeness has a significant impact on users' behavior in using artificial intelligence technology. Patil et al. showed that in terms of consumers adopting mobile payment behavior, consumers' personal innovativeness will play a positive role.

且十分重要的作用[22]。Slade 等人在英国探讨影响非用户在英国采用 RMP 意愿因素时也发现,个人创新性显著影响了非用户采用 RMP 的意愿,从而间接影响其使用 RMP 的行为[51]。基于上述分析,本研究假设:
And plays a very important role [22]. Slade et al. explored the factors influencing non-users' willingness to adopt RMP in the UK and found that individual innovativeness significantly influenced non-users' willingness to adopt RMP, thereby indirectly influencing their use of RMP [51]. Based on the above analysis, this study assumes:

: 个人创新性正向影响物流行业用户对人工智能技术的使用行为
: Personal innovativeness positively influences logistics industry users' usage behavior of artificial intelligence technology

2.2.8 人工智能焦虑和使用行为 2.2.8 Artificial Intelligence Anxiety and Usage Behavior

人工智能焦虑指的是对人工智能技术在个人或社会生活中可能引发问题的过度焦虑和恐慌[26]。Kaya 等人将人工智能焦虑归类为“工作替代焦虑”,这种焦虑通常涉及到对未来岗位的不确定性或者与自动化系统竞争等问题[52]。 Bossmann 和 Julia 在研究中表明,人工智能所导致的失业现象加剧了人与人之间不平等的问题, 进而导致人们对于采用人工智能技术产生了恐惧[25]。另外还有研究提出, 人工智能的发展可能会引发社会焦虑并带来安全问题, 从而对人们使用人工智能技术的行为造成消极影响[53]。Kwarteng 等人提到,在交通领域内人工智能焦虑会影响使用欲望, 进而影响用户未来对人工智能的使用[49]。过去文献提出, 焦虑与新技术接受之间存在关联。具体而言, 较高水平的人工智能焦虑通常与较低水平的技术接受和使用相关联[54]。基于上述分析, 本研究假设:
Artificial intelligence anxiety refers to the excessive anxiety and panic about the potential problems that artificial intelligence technology may cause in personal or social life. Kaya et al. classified artificial intelligence anxiety as "job replacement anxiety," which typically involves uncertainties about future positions or competition with automated systems. Bossmann and Julia's research shows that the unemployment caused by artificial intelligence exacerbates inequality among people, leading to fear of adopting artificial intelligence technology. Additionally, studies suggest that the development of artificial intelligence may trigger social anxiety and bring security issues, thereby negatively impacting people's behavior in using artificial intelligence technology. Kwarteng et al. mentioned that artificial intelligence anxiety in the transportation sector can affect the desire to use it, thereby influencing users' future use of artificial intelligence. Past literature has suggested a relationship between anxiety and acceptance of new technology. Specifically, higher levels of artificial intelligence anxiety are usually associated with lower levels of technology acceptance and use. Based on the above analysis, this study assumes:

H8:人工智能焦虑会对物流行业用户的人工智能技术使用行为产生负面影响。
H8: Artificial intelligence anxiety will have a negative impact on the artificial intelligence technology usage behavior of users in the logistics industry.

2.2.9 人工智能焦虑和绩效期望 2.2.9 Artificial Intelligence Anxiety and Performance Expectations

人工智能焦虑可能降低个体对人工智能技术的期望性。在物流行业中,当用户对人工智能的可靠性、安全性和稳定性产生焦虑时,他们可能会降低对人工智能系统的信任度[25], 这种信任度的降低可能会导致用户对人工智能系统的使用意愿减弱, 从而影响到系统的绩效表现。另一方面, 有些用户可能会过度依赖人工智能系统, 认为系统能够解决所有问题。然而, 当系统出现故障或不符合预期时, 用户可能会感到沮丧和失望[22], 进而影响到他们的工作绩效和满意度。一项针对自动驾驶汽车社会困境的研究表明, 人们普遍担心当自动驾驶汽车在运输乘客途中发生事故时,人工智能系统面临的抉择是牺牲乘客还是碾压行人[55]。这种对人工智能技术能力和影响的疑虑与担忧可能会让用户对人工智能贴上 “无用性” 的标签。基于上述分析,本研究假设:
Artificial intelligence anxiety may reduce individuals' expectations of artificial intelligence technology. In the logistics industry, when users have anxiety about the reliability, security, and stability of artificial intelligence, they may reduce their trust in artificial intelligence systems [25]. This decrease in trust may lead to a weakening of users' willingness to use artificial intelligence systems, thereby affecting the performance of the system. On the other hand, some users may overly rely on artificial intelligence systems, believing that the system can solve all problems. However, when the system fails or does not meet expectations, users may feel frustrated and disappointed [22], thereby affecting their work performance and satisfaction. A study on the social dilemma of autonomous driving cars suggests that people are generally concerned about the choice artificial intelligence systems face when accidents occur while transporting passengers, whether to sacrifice passengers or run over pedestrians [55]. These doubts and concerns about the capabilities and impacts of artificial intelligence technology may lead users to label artificial intelligence as "useless." Based on the above analysis, this study assumes:

H9:人工智能焦虑会对物流行业用户的人工智能技术绩效期望产生负面影响。
H9: Artificial intelligence anxiety will have a negative impact on the artificial intelligence technology performance expectations of users in the logistics industry.

2.2.10 人工智能焦虑和努力期望 2.2.10 Artificial Intelligence Anxiety and Effort Expectation
相关研究发现, 由于智能技术的复杂性和不确定性, 个体可能需要花费更多的时间和精力来学习和适应人工智能技术[22]。同时, 人工智能焦虑还可能影响人们对未来职业发展的信心。当物流行业人员面临可能被人工取代的风险时, 员工们就可能会对自己的职业前景感到迷茫和不安。这种不确定性可能导致他们缺乏明确的目标和计划, 从而降低了他们为实现目标而努力的动力。过去的研究提出, 计算机焦虑程度高的个体往往对计算机技术抱有较低的努力期望[56, 57], 而人工智能焦虑模型则是在以往技术进步焦虑模型(例如计算机焦虑)的基础上发展而来[58]。因此,我们可以推断在物流行业内,若存在着广泛的人工智能技术焦虑现象,用户对人工智能技术的努力期望很可能会持续维持在较低水平。基于上述分析, 本研究假设:
Relevant studies have found that due to the complexity and uncertainty of intelligent technology, individuals may need to spend more time and effort to learn and adapt to artificial intelligence technology. At the same time, artificial intelligence anxiety may also affect people's confidence in future career development. When logistics industry personnel face the risk of being replaced by artificial intelligence, employees may feel confused and anxious about their career prospects. This uncertainty may lead to a lack of clear goals and plans, thereby reducing their motivation to strive for their goals. Past research has suggested that individuals with high computer anxiety tend to have lower effort expectations for computer technology, and the artificial intelligence anxiety model is developed based on previous technology advancement anxiety models (such as computer anxiety). Therefore, we can infer that if there is widespread artificial intelligence technology anxiety in the logistics industry, users' effort expectations for artificial intelligence technology are likely to remain at a low level. Based on the above analysis, this study assumes:

H10:人工智能焦虑会对物流行业用户的人工智能技术努力期望产生负面影响。
H10: Artificial intelligence anxiety will have a negative impact on the expectations of artificial intelligence technology efforts of users in the logistics industry.

2.2.11 人工智能焦虑和社会影响 2.2.11 Artificial Intelligence Anxiety and Social Impact

当物流行业中广泛采用人工智能技术时,如果公众对人工智能存在焦虑情绪, 这种情绪可能会通过人际关系和社会来往等方式扩散开来, 进而影响到个体对人工智能技术的接受程度。具体来说, 如果个体周围的人或组织对人工智能持负面态度或存在担忧, 他们可能会向个体传递出消极的信息和期望, 导致个体对人工智能技术的接受度降低。Stahl 等人提出, 人工智能技术快速发展带来的伦理和技术问题可能会产生负面的社会影响[59], 同时 Rhee 也认为, 强人工智能焦虑会造成以人为本的社会价值威胁[60]。自动化和机器学习的进步意味着对物流行业中一些传统职业存在潜在风险。Acemoglu 和 Restrepo 强调,人工智能技术的引入降低了成本, 但也显著增加了每年因此失业的美国人数 [61], 引发了社会上的恐慌, 这种恐慌可能带来负面的社会影响。基于上述分析, 本研究假设:
When artificial intelligence technology is widely adopted in the logistics industry, if the public has anxiety about artificial intelligence, this emotion may spread through interpersonal relationships and social interactions, thereby affecting individuals' acceptance of artificial intelligence technology. Specifically, if the people or organizations around an individual hold negative attitudes or concerns about artificial intelligence, they may transmit negative information and expectations to the individual, leading to a decrease in the individual's acceptance of artificial intelligence technology. Stahl et al. proposed that the ethical and technical issues brought about by the rapid development of artificial intelligence technology may have negative social impacts, and Rhee also believes that anxiety about strong artificial intelligence can pose a threat to human-centered social values. The progress of automation and machine learning implies potential risks to some traditional occupations in the logistics industry. Acemoglu and Restrepo emphasize that the introduction of artificial intelligence technology reduces costs but also significantly increases the number of Americans unemployed each year as a result, causing panic in society, which may bring negative social impacts. Based on the above analysis, this study assumes:
H11:人工智能焦虑会对物流行业用户采用人工智能技术时的社会影响力产生负面影响。
H11: Artificial intelligence anxiety will have a negative impact on the social influence of logistics industry users when adopting artificial intelligence technology.
图 1 基于 UTAUT 理论的人工智能技术采用模型
Figure 1 Adoption Model of Artificial Intelligence Technology Based on UTAUT Theory

3.研究方法 3. Research Methods

3.1 问卷设计 3.1 Questionnaire Design

为验证假设并深入研究人工智能技术在物流领域的应用情况, 我们选用了问卷调查这一科学方法进行进一步的探究。问卷分为前言和正文两个部分:前言旨在向调查对象阐明调查目的和人工智能在物流行业中的当前应用状况, 旨在帮助调查对象建立对该技术的基本认知并消除疑虑,自愿参与问卷的填写。正文则由一系列问题构成,分为两个部分:第一部分主要负责收集调查对象的基本信息, 包括性别、年龄、学历、收入、职称、工作时长等; 第二部分是调查问卷的核心,经过文献综述和理论分析,我们设计了该部分的问卷结构,并确保每个结构都涵盖了 3-5 个问题以进行测量。每个问题采用 5 级克特量表,
To verify the hypothesis and further study the application of artificial intelligence technology in the field of logistics, we used a scientific method of questionnaire survey for further exploration. The questionnaire is divided into two parts: the preface aims to clarify the purpose of the survey and the current application status of artificial intelligence in the logistics industry to help the respondents establish a basic understanding of the technology and eliminate doubts, and voluntarily participate in the questionnaire. The main body consists of a series of questions, divided into two parts: the first part is mainly responsible for collecting basic information of the respondents, including gender, age, education, income, title, work duration, etc.; the second part is the core of the survey questionnaire. After literature review and theoretical analysis, we designed the questionnaire structure of this part, ensuring that each structure covers 3-5 questions for measurement. Each question uses a 5-point Likert scale.

受访者依据自身实际情况选择 1 至 5 的评分, 其中 1 代表非常不同意, 5 代表非常同意, 以获得对每个测量项的详实反馈数据。
Respondents choose a rating from 1 to 5 based on their actual situation, where 1 represents strongly disagree and 5 represents strongly agree, to obtain detailed feedback data for each measurement item.
在的正式调查之前, 我们进行了两次预测试, 以确保调查工具的准确性和可靠性,并优化数据收集流程。
Before the formal investigation, we conducted two pre-tests to ensure the accuracy and reliability of the survey tools, and optimize the data collection process.
在第一次预测试中, 我们选择了少量物流行业人员进行问卷填写, 并收集了他们的反馈和意见。我们发现部分问题表述过于复杂,难以理解,同时问卷中的某些问题顺序需要调整以提高逻辑性和流畅性。根据这些发现,我们对问卷进行了相应的修改。
In the first pre-test, we selected a small number of logistics industry personnel to fill out the questionnaire and collected their feedback and opinions. We found that some questions were too complex and difficult to understand, and that the order of some questions in the questionnaire needed to be adjusted to improve logic and fluency. Based on these findings, we made corresponding modifications to the questionnaire.
基于第一次预测试的改进, 我们进行了第二次预测试。再次邀请受访者进行问卷填写,并深入探讨了问卷的实用性和完整性。在第二次预测试中,我们进一步发现部分问题的选项设置不够全面,未能覆盖所有可能的情况,同时问卷长度较长, 可能导致填写者失去耐心。针对这些问题, 我们增加了选项设置以全面反映被调查者的实际情况,并精简了问卷长度以提高填写效率。
Based on the improvements from the first pretest, we conducted a second pretest. We once again invited respondents to fill out the questionnaire and delved into the practicality and completeness of the questionnaire. In the second pretest, we further found that the options for some questions were not comprehensive enough and did not cover all possible scenarios. Additionally, the questionnaire was too long, which could lead to respondents losing patience. To address these issues, we added more options to comprehensively reflect the actual situations of the respondents and streamlined the questionnaire length to improve efficiency in filling it out.
通过两次预测试, 我们成功地识别了问卷中存在的问题, 并采取了相应的改进措施。这些措施不仅提高了问卷的有效性和可靠性, 也为正式调查的数据收集打下了坚实的基础。
Through two pre-tests, we successfully identified the issues in the questionnaire and took corresponding improvement measures. These measures not only improved the effectiveness and reliability of the questionnaire but also laid a solid foundation for data collection in the formal survey.

3.2 数据收集与统计 3.2 Data Collection and Statistics

对于数据收集, 我们采用了有目的性的样本, 选择了物流管理和项目规划领域的专业人士, 并且这些人正在或者考虑采用人工智能技术。为了全面了解他们对于将人工智能技术应用于物流行业建设的态度和行为, 我们使用了线上和线下调查相结合的方法对其进行问卷发放。受访者被明确告知本研究仅用于学术研究, 并确保了他们的匿名性和保密性。同时, 我们还提供了一份详细的指导方针, 以确保受访者能够准确填写调查问卷, 从而提高了回复率。受访者被要求在接收电子邮件后的 20 天内回复。在 2024 年 4 月 15 日至 2024 年 5 月 5 日期间, 我们向 356 位受访者分发了调查问卷, 收到了 295 份回复, 回复率达 。在排除填写不完整、答案重复率过高(例如全文都是“非常满意”)、问卷填写时间过短等无效问卷后, 最终得到 268 份有效问卷, 有效问卷率为
For data collection, we used purposeful sampling, selecting professionals in the fields of logistics management and project planning who are currently using or considering adopting artificial intelligence technology. In order to fully understand their attitudes and behaviors towards applying artificial intelligence technology in the logistics industry, we distributed questionnaires using a combination of online and offline surveys. Respondents were explicitly informed that the study was for academic research purposes only, ensuring their anonymity and confidentiality. Additionally, we provided detailed guidelines to ensure that respondents could accurately complete the survey, thereby increasing the response rate. Respondents were asked to reply within 20 days of receiving the email. Between April 15, 2024, and May 5, 2024, we distributed surveys to 356 respondents and received 295 responses, resulting in a response rate of . After excluding incomplete responses, excessively high repetition rates (e.g., all answers being "very satisfied"), and surveys with excessively short completion times, we obtained 268 valid responses, with an effective response rate of .

4. 数据结果与分析 4. Data Results and Analysis

4.1.信度和效度 4.1. Reliability and Validity

为了估计结构的内部一致性, 计算了复合信度(CR)和 Cronbach'salpha 值。
To estimate the internal consistency of the structure, composite reliability (CR) and Cronbach's alpha values were calculated.
通过计算所有结构的复合信度(CR)来估计其内部一致性, 这是一种评估测量模型中各潜在变量可靠性的重要方法。复合信度综合考虑了因子载荷
By calculating the composite reliability (CR) of all structures to estimate their internal consistency, this is an important method for assessing the reliability of latent variables in measurement models. Composite reliability takes into account factor loadings.
(factorloadings) 的大小与潜变量之间的误差变异, 从而为研究者提供了一个关于测量工具整体稳定性和一致性的量化指标。高复合信度表明观测到的变量较好地代表了其背后的潜变量, 减少了测量误差的影响, 增强了研究结果的可信度。
The size of factor loadings is related to the error variance between latent variables, providing researchers with a quantitative indicator of the overall stability and consistency of the measurement tool. High composite reliability indicates that the observed variables represent the underlying latent variables well, reducing the impact of measurement errors and enhancing the credibility of research results.
复合信度通常由平均提取公因数载荷(AVE, AverageVarianceExtracted)和残差平方和的估计值计算得出, 公式可表示为:CR= 。当 值大于 0.7 时, 一般认为该潜变量的测量具有较好的内部一致性, 能够较为可靠地进行后续的结构模型分析。
Composite reliability is typically calculated from the estimated values of the average extracted factor loadings (AVE, Average Variance Extracted) and the sum of squared residuals, represented by the formula: CR= . When the value is greater than 0.7, it is generally considered that the measurement of the latent variable has good internal consistency, and can be reliably used for subsequent structural model analysis.
这项评估的用处在于,它帮助研究者确保所构建的测量模型不仅是理论上合理, 而且在实践中也是稳健和可靠的。在学术研究和市场调研等领域, 一个高复合信度的模型能够增强研究结论的说服力,因为这表明研究中使用的量表或问卷能够稳定且一致地测量所关注的构造。此外,这也为其他学者复现实验和比较研究结果提供了基础, 促进了知识的有效积累和科学的进步。因此, 在论文中报告并讨论复合信度的估计值, 是对研究严谨性和有效性的重要补充。
The usefulness of this assessment lies in helping researchers ensure that the measurement model they construct is not only theoretically sound, but also robust and reliable in practice. In fields such as academic research and market research, a highly composite reliable model can enhance the persuasiveness of research conclusions, as it indicates that the scales or questionnaires used in the study can consistently and stably measure the constructs of interest. Furthermore, this also provides a foundation for other scholars to replicate experiments and compare research results, promoting effective accumulation of knowledge and scientific progress. Therefore, reporting and discussing estimates of composite reliability in a paper is an important supplement to the rigor and effectiveness of the research.
如表 1 所示: As shown in Table 1:
表 1:结构信度和效度 Table 1: Structural Reliability and Validity
Constructreliabilityandvalidity
Construct reliability and validity
Cronbach's
alpha
CR AVE
AIA 0.785 0.861 0.607
EE 0.842 0.905 0.760
FC 0.856 0.912 0.776
II 0.821 0.893 0.736
PE 0.850 0.909 0.769
SI 0.858 0.913 0.778
UB 0.853 0.911 0.773
通过计算每个结构的克朗巴赫的 alpha( )来评估这些结构的一致性,这是一种常用的统计方法,用于衡量量表或问卷中各个项目间的一致性或同质性。
By calculating the Cronbach's alpha ( ) of each structure to evaluate the consistency of these structures, this is a commonly used statistical method for measuring the consistency or homogeneity between items in a scale or questionnaire.
Alpha 系数是基于项目间相关性和项目数量计算得出的,其值范围从 0 到 1 ,值越接近 1 表示各观测项之间的一致性越高, 量表的内在一致性越好, 即测量误差相对较小。Cronbach'salpha 系数是最常用的可靠性测量指标。
The Alpha coefficient is calculated based on the correlation between projects and the number of projects, with values ranging from 0 to 1. The closer the value is to 1, the higher the consistency between the observed items, the better the internal consistency of the scale, and the relatively smaller the measurement error. Cronbach's alpha coefficient is the most commonly used reliability measurement indicator.
大多数研究认为克朗巴赫 系数为 0.9 或更高表示可靠性极好, 0.8 和 0.9 之间表示可靠性良好, 0.7 和 0.8 之间表示可靠性可接受, 0.6 和 0.7 之间表示可靠性中等, 0.5 和 0.6 之间表示可靠性差, 低于 0.7 表示可靠性差。0.5 表示可靠性不可接受。
Most studies consider the Cronbach's alpha coefficient to be 0.9 or higher indicating excellent reliability, between 0.8 and 0.9 indicating good reliability, between 0.7 and 0.8 indicating acceptable reliability, between 0.6 and 0.7 indicating moderate reliability, between 0.5 and 0.6 indicating poor reliability, and below 0.5 indicating unacceptable reliability.
所有构式的 Cronbach' salpha 值在 之间, CR 值在 之间, 均超过 0.7 的阈值, 表明具有较高的内部一致性和信度。
The Cronbach's alpha values of all constructs are between , and the CR values are between , both exceeding the threshold of 0.7, indicating high internal consistency and reliability.
各潜变量 Cronbach'salpha 值介于 之间, 均大于 0.7 ; 各潜变量 Compositereliability 值介于 之间, 均大于 0.7 ; 表明各潜变量题项具有较高的内部一致性。各潜变量 AVE 值介于 0.607 0.778 之间, 均大于 0.5 ; 表明各潜变量同时解释所属全部题项的能力相当强。
Each latent variable has a Cronbach's alpha value between , all greater than 0.7; each latent variable has a Composite reliability value between , all greater than 0.7; indicating high internal consistency of the latent variable items. The AVE values of each latent variable range between 0.607 and 0.778, all greater than 0.5; indicating a strong ability of each latent variable to explain all its corresponding items.

4.2 多重共线性检验 4.2 Multicollinearity Test

计算方差膨胀因子(VIF)值以检查多重共线性问题。
Calculate the Variance Inflation Factor (VIF) value to check for multicollinearity issues.
通过估计每个结构的方差膨胀因子(VIF)来进行多重共线性检验。通过计算每个结构的克伦巴赫的 alpha( )来评估这些结构的一致性。经过所有这些计算之后, 似乎所有的参数都在可接受的范围内。
Conduct multicollinearity test by estimating the variance inflation factor (VIF) for each structure. Evaluate the consistency of these structures by calculating Cronbach's alpha ( ) for each structure. After all these calculations, it seems that all parameters are within an acceptable range.
因此, 这些项目是可靠的, 并且这些结构是一致的, 有效的, 并且不存在多重共线性缺陷。
Therefore, these projects are reliable, and these structures are consistent, effective, and free of multicollinearity defects.
如表 2 所示, 所有的 VIF 值都在 1.000 到 2.142 之间, 远低于 3 的阈值, 表明模型中没有严重的多重共线性问题。
As shown in Table 2, all VIF values are between 1.000 and 2.142, well below the threshold of 3, indicating no serious multicollinearity issues in the model.
表 2:VIF-内部模型 Table 2: VIF-Internal Model
VIF-Innermodel
VIF
AIA->EE 1.000
AIA->PE 1.215
AIA->SI 1.000
AIA->UB 1.588
EE->PE 1.206
EE->UB 2.142
FC->UB 1.404
II->PE 1.022
II->UB 1.415
PE->UB 1.991
SI->UB YES->NO 1.542
各路径的 VIF 值介于 1.000 2.142 之间,均小于 3,表明模型不存在严重的共线性问题。
The VIF values of each path range from 1.000 to 2.142, all less than 3, indicating that the model does not have a serious multicollinearity problem.

4.3 区分效度 4.3 Discriminant Validity

判别效度采用 Fornell-Larcker 标准和 HTMT 比值进行评价。
Discriminant validity is evaluated using the Fornell-Larcker criterion and the HTMT ratio.
判别效度测试(Fornell&Larcker,1981)用来确定项目是否可以完全解释自己的结构和弱与其他结构, 证实了每个构造的平方根大于相应的构造与其他结构的相关系数。
Discriminant validity test (Fornell & Larcker, 1981) is used to determine whether a construct can fully explain its own structure and is weakly related to other structures, confirming that the square root of each construct is greater than the corresponding correlation between the construct and other structures.
的平方根显示在表 3 中的对角线位置,相关系数显示在非对角线位置。
The square root of is displayed on the diagonal position in Table 3, and the correlation coefficients are displayed on the off-diagonal positions.
为了补充福内尔和拉克尔的标准,我们采用了异质单性状(HTMT)相关比率检验(Henseler 等人,2014)。结果显示, 所有构造的值都小 S0.85 (Voorheesetal.,2016)。这一结果证实了这些构造的判别有效性。
In order to supplement the standards of Fornell and Larcker, we adopted the Heterotrait-Monotrait (HTMT) ratio test (Henseler et al., 2014). The results show that all constructed values are less than 0.85 (Voorhees et al., 2016). This result confirms the discriminant validity of these constructs.
如表 3 所示, 所有构造 AVE 的平方根均大于构造之间的相关性, 说明区分效度良好。
As shown in Table 3, the square root of all the constructs' AVE is greater than the correlations between the constructs, indicating good discriminant validity.
表 3:区分效度-Fornell-Larcker 标准
Table 3: Discriminant Validity - Fornell-Larcker Standard
Discriminantvalidity-Fornell-Larckercriterion
Discriminant validity - Fornell-Larcker criterion
AIA EE FC II PE SI UB
AIA
EE -0.410
FC -0.137 0.500
II -0.135 0.105 0.146
PE -0.476 0.403 0.131 0.483
SI -0.411 0.494 0.187 0.100 0.125
UB -0.426 0.581 0.565 0.395 0.449 0.460
注: 对角线黑体字为各维度 AVE 平根
Note: The boldface diagonal letters represent the square root of the AVE of each dimension
Discriminantvalidity-HTMT
Discriminant validity-HTMT
AIA EE FC II PE SI UB
AIA
EE 0.503
FC 0.169 0.590
II 0.168 0.124 0.175
PE 0.579 0.475 0.153 0.577
SI 0.498 0.579 0.216 0.117 0.145
UB 0.519 0.685 0.662 0.471 0.525 0.536
各维度 AVE 平方根均大于其与其他维度间相关系数绝对值, 表明数据具有
The square root of the Average Variance Extracted (AVE) for each dimension is greater than the absolute value of the correlation coefficients between it and other dimensions, indicating that the data is

区别效度。此外, 我们进一步通过 HTMT 指标来验证数据的区别效度。当 HTMT 值落在 0.117 至 0.685 的范围内时, 通常被视为数据具有区别效度的可靠指标。在本文中,如表 4 所示,所有计算的 HTMT 值均低于 0.85 的阈值,这一结果进一步强化了构念之间区分效度的可靠性。
Discriminant validity. In addition, we further validate the discriminant validity of the data through the HTMT index. When the HTMT value falls within the range of 0.117 to 0.685, it is generally considered a reliable indicator of discriminant validity of the data. In this study, as shown in Table 4, all calculated HTMT values are below the threshold of 0.85, further reinforcing the reliability of discriminant validity between concepts.

4.4 路径系数 4.4 Path Coefficient

结构方程模型(SEM)作为统计框架来深入研究变量之间复杂的关系。在研究中, SEM 通常用于验证理论模型或探索变量之间的因果关系。以下是对其数学背景的简洁解释:
Structural Equation Modeling (SEM) is a statistical framework used to investigate complex relationships between variables. In research, SEM is commonly used to validate theoretical models or explore causal relationships between variables. Here is a concise explanation of its mathematical background:
变量表示:在 SEM 中,变量分为观察变量(显变量)和潜在变量。观测变量是可以直接测量的, 而潜变量则不能直接观测, 但可以通过观测变量的测量间接推断。
Variable representation: In SEM, variables are divided into observed variables (manifest variables) and latent variables. Observed variables can be directly measured, while latent variables cannot be directly observed, but can be indirectly inferred through the measurement of observed variables.
鉴于本文的研究假设,观测变量如下:PE、EE、AIA、II、SI、FC、UB
Given the research hypothesis of this article, the observed variables are as follows: PE, EE, AIA, II, SI, FC, UB
潜在变量是: UB、PE、EE、SI The latent variables are: UB, PE, EE, SI
路径模型:SEM 采用路径模型来描述变量之间的关系。路径模型利用路径系数来量化变量之间的直接或间接关系。这些路径系数说明了一个变量的变化如何影响其他变量。
Path model: SEM uses path models to describe the relationships between variables. Path models use path coefficients to quantify the direct or indirect relationships between variables. These path coefficients explain how changes in one variable affect other variables.
测量模型:SEM 还涉及测量模型,它解释观察变量和潜在变量之间的关系。测量模型通过因子载荷量化观察变量和潜在变量之间的关联。
Measurement model: SEM also involves a measurement model, which explains the relationship between observed variables and latent variables. The measurement model quantifies the relationship between observed variables and latent variables through factor loadings.
模型拟合:SEM 中的模型拟合指数用于评估模型与观测数据之间的拟合度。常见的模型拟合指标包括 x 2 检验、RMSEA、CFI 等。
Model Fit: Model fit indices in SEM are used to evaluate the fit between the model and the observed data. Common model fit indices include x2 test, RMSEA, CFI, etc.
在研究中, SEM 用于验证研究假设并探索变量之间的关系。通过适当的模型构建和参数估计, SEM 帮助研究人员理解变量之间的复杂关系, 从而为研究提供强有力的支持和解释。
In research, SEM is used to validate research hypotheses and explore relationships between variables. Through appropriate model construction and parameter estimation, SEM helps researchers understand the complex relationships between variables, providing strong support and explanations for research.
SEM 模型方程通常包括测量模型和结构模型。 SEM model equations typically include measurement models and structural models.
测量模型: 测量模型描述了观测变量和潜在变量之间的关系, 通常由因子载荷表示。测量模型方程可以表示为:
Measurement Model: The measurement model describes the relationship between observed variables and latent variables, usually represented by factor loadings. The measurement model equation can be represented as:
在这里: Here:
是观测变量的向量;  is a vector of observed variables;
为因子载荷矩阵, 表示观测变量与潜变量之间的线性关系;
is the factor loading matrix, representing the linear relationship between observed variables and latent variables;
是潜在变量的向量;  is a vector of latent variables;
是测量误差的向量。  is the vector of measurement errors.
结构模型:结构模型描述了潜在变量之间的关系,通常由路径系数表示。结构模型方程可表示为:
Structural model: The structural model describes the relationships between latent variables, usually represented by path coefficients. The structural model equation can be expressed as:
在这里: Here:
是潜在变量的向量;  is a vector of latent variables;
是路径系数矩阵,表示潜变量之间的直接或间接关系;
is the path coefficient matrix, representing the direct or indirect relationships between latent variables;
是结构错误的向量。  is a vector with a structural error.
通过将测量模型和结构模型相结合,可以构建全面的 SEM 模型方程组,以分析变量之间的复杂关系并验证研究假设。
By combining measurement models and structural models, a comprehensive SEM model equation system can be constructed to analyze the complex relationships between variables and validate research hypotheses.
结构方程模型(SEM)作为一种强大的统计工具的应用在研究工作中具有重要意义。首先, 选择 SEM 模型的理由之一在于它能够解决复杂的研究问题和多个变量之间的关系。在物流行业, 众多潜在影响因素和变量相互作用, 传统的统计方法可能无法充分捕捉变量之间错综复杂的关系。SEM 能够同时考虑观察变量和潜在变量,从而有助于对变量之间的关系进行更全面的分析。
Structural Equation Modeling (SEM) as a powerful statistical tool is of great significance in research work. One of the reasons for choosing the SEM model is that it can address complex research questions and relationships among multiple variables. In the logistics industry, numerous potential influencing factors and variables interact with each other, and traditional statistical methods may not fully capture the intricate relationships among variables. SEM can simultaneously consider observed variables and latent variables, thereby aiding in a more comprehensive analysis of relationships among variables.
此外, SEM 提供了一个用于验证理论模型或研究假设的结构化框架。在物流行业, 研究人员可能会关注绩效预期、努力预期、社会影响力等各个方面对人工智能(AI)技术采用行为的影响。通过 SEM,研究人员可以构建包含多个潜在变量和观察变量的复杂模型, 并验证这些变量之间的关系,从而更深入地了解人工智能技术采用行为的潜在机制。
In addition, SEM provides a structured framework for validating theoretical models or researching hypotheses. In the logistics industry, researchers may focus on the impact of various aspects such as performance expectations, effort expectations, and social influence on the adoption behavior of artificial intelligence (AI) technology. Through SEM, researchers can construct complex models containing multiple latent and observed variables, and validate the relationships between these variables, thereby gaining a deeper understanding of the potential mechanisms of AI technology adoption behavior.
此外, SEM 还可以进行模型比较和路径分析, 帮助研究人员确定最佳模型和路径结构。在研究中, 模型和路径的选择可以显着影响最终的研究结果和结论。通过 SEM, 研究人员可以根据实际数据比较不同的模型, 选择最能解释数据的模型, 从而提高研究的可信度和可解释性。
In addition, SEM can also perform model comparison and path analysis, helping researchers determine the best model and path structure. In research, the choice of models and paths can significantly impact the final research results and conclusions. Through SEM, researchers can compare different models based on actual data, select the model that best explains the data, thereby enhancing the credibility and interpretability of the research.
对于人工智能技术在物流行业的采用行为的研究, SEM 作为一种灵活而强大的统计工具, 能够有效分析复杂的多变量关系, 验证研究假设, 并提供结构化框架, 从而有助于更深入地理解物流行业的研究问题。这个领域。
Research on the adoption of artificial intelligence technology in the logistics industry, SEM, as a flexible and powerful statistical tool, can effectively analyze complex multivariate relationships, validate research hypotheses, and provide a structured framework, thereby helping to gain a deeper understanding of research issues in the logistics industry.

接下来, 研究将利用 AMOS 软件构建结构方程模型。
Next, the study will use AMOS software to build a structural equation model.
通过分析路径系数来确定构造之间的直接影响。如表 5 所示, 所有路径的 值都小于 0.05 , 说明变量之间存在显著的关系。
By analyzing the path coefficients to determine the direct effects between constructs. As shown in Table 5, all path coefficients (p-values) are less than 0.05, indicating a significant relationship between variables.
Analysis of path coefficients in structural equation modeling
Path Hypothesis
Originalsample

Standard deviation (STDEV)
Standarddeviation
(STDEV)
Tstatistics
-value Remarks
PE->UB H1 0.153 0.062 2.492 0.013 Supported
EE->UB H2 0.144 0.06 2.398 0.017 Supported
EE->PE H3 0.226 0.054 4.219 0 Supported
SI->UB YES->NO H4 0.227 0.048 4.698 0 Supported
FC->UB H5 0.383 0.045 8.539 0 Supported
II->PE H6 0.415 0.047 8.822 0 Supported
II->UB H7 0.211 0.054 3.874 0 Supported
AIA->UB H8 -0.12 0.055 2.159 0.031 Supported
AIA->PE H9 -0.327 0.048 6.803 0 Supported
AIA->EE H10 -0.41 0.049 8.445 0 Supported
AIA->SI H11 -0.411 0.054 7.639 0 Supported
路径 PE->UB 回归系数为 , 置信区间为 ,不包含 0 , 因此, 路径 PE->UB 是显著的。
The regression coefficient of path PE->UB is , with a confidence interval of , not including 0, therefore, the path PE->UB is significant.

路径 EE->UB 回归系数为 , 置信区间为 ,不包含 0 , 因此, 路径 EE->UB 是显著的。
The regression coefficient of path EE->UB is , with a confidence interval of , not including 0, therefore, the path EE->UB is significant.
路径 EE->PE 回归系数为 , 置信区间为 , 不包含 0 , 因此, 路径 EE->PE 是显著的。
The return coefficient of path EE->PE is , with a confidence interval of , not including 0, therefore, the path EE->PE is significant.
路径 SI->UB 回归系数为 , 置信区间为 , 不包含 0 , 因此, 路径 SI->UB 是显著的。
The regression coefficient of path SI->UB is , with a confidence interval of , not including 0, therefore, the path SI->UB is significant.
路径 FC->UB 回归系数为 , 置信区间为 , 不包含 0 , 因此, 路径 FC->UB 是显著的。
The regression coefficient of path FC->UB is , with a confidence interval of , not including 0, therefore, the path FC->UB is significant.
路径 II->PE 回归系数为 , 置信区间为 , 不包含 0 , 因此, 路径 II->PE 是显著的。
The return coefficient of path II->PE is , with a confidence interval of , not including 0, therefore, the path II->PE is significant.
路径 II->UB 回归系数为 , 置信区间为[0.106,0.321], 不包含 0 , 因此, 路径 II->UB 是显著的。
The regression coefficient of path II->UB is , with a confidence interval of [0.106, 0.321], not including 0, therefore, the path II->UB is significant.
路径 AIA->UB 回归系数为 , 置信区间为 ,不包含 0 , 因此, 路径 AIA->UB 是显著的。
The return coefficient of path AIA->UB is , with a confidence interval of , not including 0, therefore, the path AIA->UB is significant.
路径 AIA->PE 回归系数为 , 置信区间为 , 不包含 0 , 因此, 路径 AIA->PE 是显著的。
The regression coefficient of path AIA->PE is , with a confidence interval of , not including 0, therefore, the path AIA->PE is significant.
路径 AIA->EE 回归系数为 , 置信区间为[-0.503,-0.314], 不包含 0 , 因此, 路径 AIA->EE 是显著的。
The regression coefficient of path AIA->EE is , with a confidence interval of [-0.503, -0.314], not including 0, therefore, the path AIA->EE is significant.
路径 AIA->SI 回归系数为 , 置信区间为[-0.515,-0.304], 不包含 0 , 因此, 路径 AIA->SI 是显著的。
The regression coefficient of path AIA->SI is , with a confidence interval of [-0.515, -0.304], not including 0, therefore, the path AIA->SI is significant.

4.5 特定的间接影响 4.5 Specific Indirect Effects

通过分析间接效应来确定变量间的中介效应。如表 6 所示,所有间接路径的 值均显著, 表明存在中介效应。
By analyzing the indirect effects to determine the mediating effects between variables. As shown in Table 6, the values of all indirect paths are significant, indicating the presence of mediating effects.
表 6:具体间接影响 Table 6: Specific Indirect Effects
Specificindirecteffects Specific indirect effects
IndirectPat
 Original sample (O)
Originalsam
ple(O)

Standard deviation (STDEV)
Standarddeviatio
n(STDEV)
Tstatistics(|O/
STDEV|)
Pvalues
AIA->EE-
PE->UB
-0.014 0.007 2.030 0.042 -0.030 -0.002
II->PE-
UB
0.064 0.027 2.344 0.019 0.013 0.120
AIA->SI-
-0.093 0.022 4.164 0.000 -0.140 -0.052
AIA->EE- -0.059 0.026 2.279 0.023 -0.113 -0.011
间接路径 AIA->EE->PE->UB 中介效应值为 , 置信区间为[-0.03,-0.002], 不包含 0 , 因此, 间接路径 AIA->EE->PE->UB 是显著的。
Indirect path AIA->EE->PE->UB has a mediation effect value of , with a confidence interval of [-0.03, -0.002], not including 0, therefore, the indirect path AIA->EE->PE->UB is significant.
间接路径 II->PE->UB 中介效应值为 , 置信区间为 [0.013,0.12], 不包含 0, 因此, 间接路径 II->PE->UB 是显著的。
Indirect path II->PE->UB, the value of the mediating effect is , with a confidence interval of [0.013,0.12], not including 0, therefore, the indirect path II->PE->UB is significant.
间接路径 AIA->SI->UB 中介效应值为 , 置信区间为 , 不包含 0 , 因此, 间接路径 AIA->SI->UB 是显著的。
The indirect effect value of the indirect path AIA->SI->UB is , with a confidence interval of , not including 0, therefore, the indirect path AIA->SI->UB is significant.
间接路径 AIA->EE->UB 中介效应值为 , 置信区间为 0.113,-0.011], 不包含 0 , 因此, 间接路径 AIA->EE->UB 是显著的。
Indirect path AIA->EE->UB has a mediation effect value of , with a confidence interval of [0.113, -0.011], not including 0, therefore, the indirect path AIA->EE->UB is significant.
间接路径 AIA->PE->UB 中介效应值为-0.05, , 置信区间为 0.096,-0.009], 不包含 0 , 因此, 间接路径 AIA->PE->UB 是显著的。
Indirect path AIA->PE->UB has a mediation effect value of -0.05, [0.096,-0.009], not including 0, therefore, the indirect path AIA->PE->UB is significant.
调查结果摘要 Summary of Investigation Results
1.测量模型信度和效度:各构式具有较高的信度和效度, 所有的 Cronbach' salpha 和复合信度值均超过 0.7 的阈值。判别效度通过 Fornell-Larcker 判据和 HTMT 比值确定。
1. Measure the reliability and validity of the model: Each construct has high reliability and validity, with all Cronbach's alpha and composite reliability values exceeding the threshold of 0.7. Discriminant validity is determined through the Fornell-Larcker criterion and the HTMT ratio.
2.多重共线性:未见严重多重共线性, VIF 值均小于 3。
2. Multicollinearity: No serious multicollinearity was observed, with VIF values all less than 3.
3.结构模型: 路径系数显示变量之间的显著关系, 支持大多数假设。间接效应分析显示中介效应显著。
3. Structural model: Path coefficients show significant relationships between variables, supporting most hypotheses. Indirect effects analysis shows significant mediating effects.

5.讨论 5. Discussion

本研究基于扩展的统一技术接受与使用理论(UTAUT)模型, 深入探讨了物流行业中员工个体层面对人工智能技术的采纳行为。该模型包含七个观测变量:绩效期望、努力期望、社会影响、促进条件、人工智能焦虑、个人创新性及使用行为, 以及四个潜在变量:使用行为、绩效期望、努力期望和社会影响。在验证性因子分析阶段,我们对数据进行了多重检验,包括信度、效度、多重共线性和区别效度检验。结果表明,使用行为与 UTAUT 模型之间存在良好的模型兼容性, 这为样本数据的结构方程分析提供了有力支持。通过构建结构方程模型验证 11 个假设(H1 至 H11), 数据处理结果显示所有假设均成立, 为人工智能背景下物流行业采用人工智能技术的影响因素提供了进一步的支持。尽管已有研究在人工智能背景下基于 UTAUT 模型对物流行业进行了分析[21],但
This study, based on the extended Unified Theory of Acceptance and Use of Technology (UTAUT) model, delves into the adoption behavior of individual employees in the logistics industry towards artificial intelligence technology. The model consists of seven observed variables: performance expectancy, effort expectancy, social influence, facilitating conditions, artificial intelligence anxiety, individual innovativeness, and usage behavior, as well as four latent variables: usage behavior, performance expectancy, effort expectancy, and social influence. In the confirmatory factor analysis stage, we conducted multiple tests on the data, including reliability, validity, multicollinearity, and discriminant validity tests. The results indicate good model compatibility between usage behavior and the UTAUT model, providing strong support for the structural equation analysis of the sample data. By constructing a structural equation model to validate 11 hypotheses (H1 to H11), the data processing results show that all hypotheses are valid, further supporting the factors influencing the adoption of artificial intelligence technology in the logistics industry under the background of artificial intelligence. Although previous studies have analyzed the logistics industry under the background of artificial intelligence based on the UTAUT model, this study provides further support for the factors influencing the adoption of artificial intelligence technology in the logistics industry.

鲜有研究将人工智能焦虑或个人创新性作为变量纳入模型当中, 且未从微观层面深入探究各变量间的相互影响关系。因此, 本研究通过扩展 UTAUT 模型,填补了这一研究空白, 并深入解释了物流行业用户对人工智能技术采纳行为的接受程度。
Rarely have studies incorporated artificial intelligence anxiety or individual innovativeness as variables into models, and have not deeply explored the interrelationships between variables at the micro level. Therefore, this study fills this research gap by extending the UTAUT model and provides a detailed explanation of the extent to which users in the logistics industry accept artificial intelligence technology adoption behavior.
现有的研究表明, 绩效期望是用户采纳人工智能技术的重要预测变量, 这些研究不仅揭示了用户对于人工智能技术高效益性的期望, 而且进一步证实了人工智能高效率工作表现是驱动用户采纳和使用行为的核心因素[3]。然而, 尽管在 的水平上发现 与使用行为之间的关系显著, 但是其影响却弱于预期值。数据处理结果显示 PE 仅以 0.153 的系数正向影响着使用行为, 在影响使用行为的五个因素中仅仅成为第四强的关系,这一发现与 Kim 等人[62]提供的 PE 与使用行为发现呈现显著分歧, 他们分析所得的结果显示绩效期望是最有影响力的因素, 这一差异可归因于运输行业内普遍存在的对人工智能技术的焦虑情绪。这一现象表明, 尽管绩效期望在学术研究及实践中占据重要地位,但在当前研究领域中,其对实际使用行为的正向推动作用却显得相对有限。
Existing research shows that performance expectations are important predictive variables for users to adopt artificial intelligence technology. These studies not only reveal users' expectations of high efficiency from artificial intelligence technology but also further confirm that the high efficiency performance of artificial intelligence is a core factor driving user adoption and usage behavior. However, although the relationship between performance expectations and usage behavior is significant at the level of , its impact is weaker than expected. Data processing results show that performance expectations positively influence usage behavior with a coefficient of only 0.153, making it the fourth strongest relationship among the five factors influencing usage behavior. This finding is in contrast to the significant difference in the relationship between performance expectations and usage behavior found by Kim et al. Their analysis results show that performance expectations are the most influential factor, which can be attributed to the anxiety towards artificial intelligence technology prevalent in the transportation industry. This phenomenon indicates that although performance expectations play an important role in academic research and practice, their positive driving force on actual usage behavior appears relatively limited in the current research field.
以前的研究相信[31], 如果用户认为人工智能技术易于使用, 他们就可以毫不费力地采用这项新技术。与绩效期望类似, 数据结果表明努力预期与人工智能技术使用行为之间的正相关关系较弱, 其路径系数仅以 0.144 影响使用行为,成为五个因素中最不具影响力的因子。尽管所需的努力会影响他们对人工智能的接受程度, 但这一因素并不是他们采纳人工智能技术的重要决定因素。
Previous research believed [31], if users perceive artificial intelligence technology as easy to use, they can effortlessly adopt this new technology. Similar to performance expectations, data results show a weak positive correlation between effort expectancy and artificial intelligence technology usage behavior, with a path coefficient of only 0.144 influencing usage behavior, becoming the least influential factor among the five factors. Although the required effort will affect their acceptance of artificial intelligence, this factor is not a key determinant of their adoption of artificial intelligence technology.
除了绩效期望与努力期望之外, 其他三个结构, 如社会影响、便利条件与个人创新性也成为了预测物流行业用户人工智能技术采纳行为的重要因素。在前面五个因素当中,便利条件成为人工智能技术使用行为的最强预测因子,其路径值为 0.383 。这一发现表明, 组织在培训和基础设施方面所提供的支持对于员工使用行为具有显著影响。此结果与先前的研究相一致, Lanhui 等人[13]的研究表明, 当用户认为他们拥有足够的资源和支持以运用人工智能技术时, 他们更倾向于接受并采纳这些技术。因此,本研究认为,更好的资源、技术和制度基础设施可以增加用户采纳人工智能技术的使用行为。
In addition to performance expectations and effort expectations, the other three constructs, such as social influence, facilitating conditions, and individual innovativeness, have also become important factors in predicting the adoption behavior of artificial intelligence technology by logistics industry users. Among the first five factors, facilitating conditions emerged as the strongest predictor of artificial intelligence technology usage behavior, with a path value of 0.383. This finding suggests that the support provided by organizations in terms of training and infrastructure significantly influences employee usage behavior. This result is consistent with previous research, as Lanhui et al.'s study [13] indicates that when users believe they have sufficient resources and support to utilize artificial intelligence technology, they are more likely to accept and adopt these technologies. Therefore, this study suggests that better resources, technology, and institutional infrastructure can increase user adoption behavior of artificial intelligence technology.
此外, 本研究的分析结果进一步揭示, 社会影响与人工智能技术使用行为之间存在显著的正相关关系, 其影响力仅次于便利条件, 路径系数为 0.227 。这
In addition, the analysis results of this study further reveal that there is a significant positive correlation between social influence and the use of artificial intelligence technology, with its influence second only to convenience conditions, and the path coefficient is 0.227.

一发现与 Dwivedi 等人[39]先前的研究结果相吻合, 均强调了社会影响在技术采纳行为中的重要性。鉴于人工智能技术在物流行业中属于新兴技术范畴, 用户可能由于缺乏充分的专业知识或判断能力来评估使用此类技术的适宜性。因此,他们往往倾向于参考其所属社会群体的态度和观点, 进而形成对使用人工智能技术的意愿和具体行为。
Once discovered, it is consistent with the previous research results of Dwivedi et al. [39], all of which emphasize the importance of social influence in technology adoption behavior. Considering that artificial intelligence technology belongs to the emerging technology category in the logistics industry, users may lack sufficient professional knowledge or judgment to assess the suitability of using such technology. Therefore, they tend to refer to the attitudes and opinions of their social groups, thereby forming willingness and specific behaviors to use artificial intelligence technology.
个人创新性被证实为一个显著的正向预测因子, 具体表现为 的路径值, 因为人工智能技术在物流行业中尚属相对较新的应用,且其操作模式与传统物流流程存在显著差。现有的大量研究已经证明个人创新性对人工智能技术使用行为有重大积极影响, 例如 Patil[22]、Hong[47]、 Cudjoe[27]等人的研究。然而, 聚焦于物流行业用户个人创新性的研究尚显不足。本研究的结果与先前文献的结论相吻合, 进一步证实了个人创新性在物流行业用户接纳人工智能技术中的核心作用。这表明, 具备较高个人创新能力的物流行业用户,更有可能展现出对人工智能技术更为积极的使用行为。
Personal innovativeness has been confirmed as a significant positive predictor, specifically manifested as the path value of