摘要：在地缘政治紧张加剧与全球能源供应链脆弱性上升的背景下，人工智能在能源安全治理中的作用日益成为学术研究和政策制定的焦点。本文基于跨国面板数据，从全球视角实证检验人工智能对能源安全风险的影响。研究结果表明：人工智能显著降低能源安全风险，其作用机制主要体现在优化能源结构和提高能源效率。异质性分析表明，人工智能对能源安全风险的影响在不同收入水平、能源依赖程度和风险环境下存在结构性差异，仅在高收入、低能源依赖和低风险国家表现出显著的能源安全风险缓释作用，而在中低收入及高风险国家可能反而加剧能源安全风险。自然资源租金与数字鸿沟对人工智能和能源安全的关系存在非线性门槛效应，前者呈现出随着自然资源租金增加人工智能对能源安全风险的抑制作用先变大后变小，后者表现为缩小数字鸿沟有助于强化人工智能抑制风险作用。上述发现揭示了人工智能在能源安全治理中的多维度与复杂性，为不同发展水平国家制定差异化政策提供了理论依据，各国应统筹推进人工智能赋能与制度能力建设，缩小数字鸿沟、优化资源配置结构，实施分层分类政策路径，以提升人工智能在不同发展阶段下的能源安全风险治理绩效。
关键词：能源安全风险；人工智能；跨国面板；非线性路径
Artificial Intelligence and Energy Security Risk: A Multidimensional Analysis Based on Cross-National Panel Data
Wang Xiaowei, Li Kai
Abstract: Against the backdrop of intensified geopolitical tensions and rising vulnerability in the global energy supply chain, the role of artificial intelligence in energy security governance has increasingly become a focus of academic research and policymaking. Based on cross-national panel data, this paper empirically examines the impact of artificial intelligence on energy security risk from a global perspective. The results show that: artificial intelligence significantly reduces energy security risk, primarily by optimizing the energy structure and enhancing energy efficiency. Heterogeneity analysis reveals that the impact of artificial intelligence on energy security risk exhibits structural differences across income levels, energy dependence, and risk environments. It demonstrates a significant mitigation effect on energy security risk only in high-income, low-energy-dependence, and low-risk countries, while potentially exacerbating energy security risk in lower-middle-income and high-risk countries. The relationship between natural resource rent and the digital divide on artificial intelligence and energy security exhibits non-linear threshold effect. The former shows that the inhibitory effect of artificial intelligence on energy security risk initially increases and then decreases with increasing natural resource rent, while the latter demonstrates that narrowing the digital divide helps strengthen the risk-mitigating effect of artificial intelligence. The above findings reveal the multidimensionality and complexity of artificial intelligence in energy security governance, providing a theoretical basis for countries at different levels of development to formulate differentiated policies. Countries should coordinate the promotion of artificial intelligence empowerment and institutional capacity building, narrow the digital divide, optimize resource allocation structure, and implement tiered and classified policy paths to improve the performance of artificial intelligence in energy security risk governance at different stages of development.
Keywords: Energy Security Risk; Artificial Intelligence; Cross-national Panel; Nonlinear Path
全文：
人工智能与能源安全风险：跨国面板视角下的多维效应识别.pdf