人工智能算法的可专利性问题研究

Research on the Patentability of Artificial Intelligence Algorithms

摘要: 人工智能算法的可专利性判定是数字时代知识产权法的核心挑战。为厘清其专利适格性边界，回应技术问题边界模糊、硬件关联标准不一及领域耦合阈值不明等难题，文章以英国Emotional Perception AI案为切入点，结合国内司法与专利审查实践，构建一个包含“技术问题识别”“硬件结合”及“领域耦合”的三维框架，并尝试为每个维度提供更具体的判断思路。研究结果表明：其一，提出技术问题的多维度判断标准，避免因涉及主观性而直接否定技术性；其二，技术手段应与硬件深度关联(如优化内存或运算效率)，体现自然规律的应用；其三，算法需突破应用领域的僵化限制，通过场景融合(如工业识别、医疗诊断)确立技术贡献。最终得出结论，即弱人工智能时代需以动态审查标准平衡创新激励与制度约束，为专利法适应技术变革提供理论支持。

Abstract: The determination of the patentability of artificial intelligence algorithms is a core challenge in intellectual property law in the digital age. To clarify the boundaries of their patent eligibility and address such challenges as the ambiguous boundaries of technical issues, inconsistent standards for hardware association, and unclear thresholds for domain coupling, this paper takes the UK Emotional Perception AI case as the starting point; combining domestic judicial and patent examination practices, it constructs a three-dimensional framework encompassing technical issue identification, hardware integration, and domain coupling. The research findings indicate that: 1. It proposes multi-dimensional criteria for judging technical issues to avoid directly denying their technical nature due to the involvement of subjectivity; 2. Technical means should be deeply associated with hardware (e.g., optimizing memory or computational efficiency) to reflect the application of natural laws; 3. Algorithms need to break through rigid restrictions in application fields and establish technical contributions through scenario integration (e.g., industrial identification, medical diagnosis). Finally, it concludes that in the era of weak artificial intelligence, dynamic examination standards are needed to balance innovation incentives and institutional constraints, providing theoretical support for the adaptation of patent law to technological changes.