人工智能司法适用中事实认定的双重困境与纾解路径

Dual Dilemma and Relieving Path of Factual Determination in the Judicial Application of Artificial Intelligence

人工智能的司法适用在事实领域面临诸多问题,其根源在于现阶段人工智能尚未达到真正意义的强人工智能。目前,人工智能的技术原理仍是深度学习,其运行模式主要是对海量数据的收集和处理。虽然生成式人工智能的长文本生成能力提升,但仍具有较强的局限性。鉴于深度学习的理论本质,人工智能在司法适用中表现出机械性特征。就人工智能在事实认定领域的具体应用而言,其面临双重困境:一方面,受自然语言技术和图像识别技术的制约,基于辩论原则的诉讼资料信息获取并不充分,不同种类的证据资料信息获取也不充分;另一方面,人工智能无法复制法官事实认定的方法,人工智能在事实认定领域既无法运用自由心证对争议事实形成内心确信,也无法适用证明责任理论分配风险进行辅助事实认定。面对上述双重困境,人工智能可在事实争议不大、权利义务关系清楚时借鉴要素式审判,提取主要事实;也可通过建构多层级的事实认定模型,拓展人工智能在事实认定领域的适用范围。

Fact finding faces many problems in the judicial application of Artificial Intelligence(AI),the root cause of which is that AI has not yet reached the stage of strong artificial intelligence in the true sense.At present,the technical principle of AI is still Deep Learning(DL),and its operation mode is mainly the collection and processing of massive data.Although the ability to generate written text of generative artificial intelligence has improved,it still has limitations.In view of the theoretical nature of Deep Learning,AI shows mechanical characteristics in judicial application.As far as the specific application of AI in the field of fact finding is concerned,it faces a double dilemma.On one hand,restricted by natural language technology and image recognition technology,the acquisition of litigation materials and information based on the principle of debate is not sufficient,and the acquisition of different types of evidence materials is also insufficient.On the other hand,AI cannot replicate the judge's method of finding facts.In the field of fact finding,AI can neither use free evidence to form inner convictions on disputed facts,nor can it apply the burden of proof theory to allocate risks to assist fact finding.Faced with the above-mentioned dual dilemma,AI can draw on the element-based trial to extract the main facts when the facts are not disputed and the relationship between rights and obligations is clear.In addition,by constructing a multi-level fact-finding model,the scope of application of AI in the field of fact-finding can be expanded.