Abstract
Purpose. To identify multidimensional barriers to AI adoption in consulting, propose breakthrough paths, and explore trust reconstruction from “tool trust” to “system trust.”
Theoretical Framework. Grounded in TAM, UTAUT, TOE, Service-Dominant Logic, and trust repair theory, this study constructs a technology-market-ethics barrier model with eight hypotheses examining barrier interactions, their effects on tool trust and system trust, and subsequent adoption intention.
Methodology. An explanatory sequential mixed-methods design was used. Quantitative data came from 212 valid surveys of consulting practitioners, analyzed via SEM. Qualitative data came from in-depth interviews with 18 industry experts, analyzed using thematic analysis.
Findings. All eight hypotheses were supported. The study reveals that technological barriers are the root cause, significantly and positively impacting market and ethical barriers. Market barriers, in turn, positively affect ethical barriers. All three barriers negatively impact tool trust. Crucially, tool trust positively influences system trust, which then positively drives adoption intention. Ethical barriers emerged as the most salient concern, scoring the highest mean. A notable "high intention, low trust" paradox was discovered, where adoption intention significantly outpaced both tool and system trust. Qualitative findings identified seven core themes, including data quality, algorithm black-box, the inimitability of tacit knowledge, client cognitive biases, resource constraints of SMEs, data privacy, and responsibility ambiguity. Breakthrough paths identified include local deployment, human-AI double review mechanisms, explainable AI (XAI) tools, accountability frameworks, and organizational training. Trust repair strategies center on transparent communication, traceable decision-making, and rapid iteration.
Conclusion and Recommendations. Technological barriers are the root cause of AI adoption obstacles in consulting, while ethical barriers are the most salient. The three barriers interact as a complex system with a “technology → market → ethics” transmission chain. Synergy between technological innovation and institutional design is essential for breakthrough. Trust evolves from tool to system trust; transparent communication and rapid response are core repair mechanisms. The “high intention, low trust” paradox highlights AI anthropomorphism as a coping mechanism. Findings extend TAM, TOE, and trust theories to the AI-consulting context, offering practical guidance for consulting firms, AI suppliers, and policymakers.
目的. 系统识别咨询行业AI技术采纳的多维障碍，提出突破路径，探索信任从“工具信任”向“系统信任”演化的动态机制。
理论框架. 基于TAM、UTAUT、TOE框架、服务主导逻辑及信任修复理论，构建技术—市场—伦理三维障碍交互模型，提出八个假设，检验障碍交互作用及其对工具信任、系统信任和采纳意愿的影响。
方法论. 采用解释性序贯混合方法。定量研究收集212份咨询从业者有效问卷，运用结构方程模型分析；定性研究对18位行业专家深度访谈，采用主题分析法。
发现. 研究提出的八个假设均获得支持。结果显示，技术障碍是根源性因素，对市场障碍和伦理障碍具有显著正向影响；市场障碍亦正向影响伦理障碍。三类障碍均负向影响工具信任。工具信任正向影响系统信任，而系统信任最终正向驱动采纳意愿。其中，伦理障碍的均值最高，成为最受关注的维度。研究还发现了一个显著的“高意愿、低信任”悖论，即从业者的采纳意愿远高于其对AI的工具信任和系统信任。定性研究提炼出七大核心主题，涵盖数据质量、算法黑箱、隐性知识难以编码、客户认知偏差、中小机构资源约束、数据隐私及责任归属模糊等。突破路径包括本地部署、人机双复核机制、可解释AI、问责机制及组织培训。信任修复策略则聚焦于透明沟通、可追溯决策与快速迭代。
结论与建议. 技术障碍是AI采纳障碍系统的根源，伦理障碍最为凸显。三类障碍形成“技术→市场→伦理”传导链条。技术创新与制度设计协同是破除壁垒的关键。信任从工具信任向系统信任演化，透明沟通与快速响应是核心修复策略。“高意愿、低信任”悖论揭示了AI拟人化作为心理调节机制。研究结果拓展了TAM、TOE及信任理论在AI咨询情境的应用，为咨询机构、AI供应商及政策制定者提供实践指导。