人工智能参与司法鉴定的实用价值、实践困境与完善建议

Practical Value, Practical Dilemmas, and Improvement Suggestions for the Involvement of Artificial Intelligence in Forensic Appraisal

摘要: 随着人工智能在司法领域的广泛应用，由人工智能参与生成的鉴定意见在诉讼中频繁出现。一方面，人工智能的应用显著提升了司法鉴定的效率和质量，在司法领域展现出巨大的潜力，但另一方面，人工智能在赋能司法鉴定的同时也在技术和法律层面面临困境：算法偏见和技术黑箱导致鉴定意见可靠性和可解释性不足，并存在数据隐私风险；法律规范滞后，人工智能鉴定主体地位和责任归属不明以及现行的程序规则和质证机制也无法有效应对人工智能鉴定意见。为此，需从规范算法数据、披露鉴定所用算法和数据、完善庭审质证规则和构建责任分担机制四个方面进行完善，促进人工智能参与司法鉴定的良性发展。

Abstract: With the extensive application of artificial intelligence (AI) in the judicial sector, expert opinions generated with the involvement of AI are increasingly prevalent in litigation. While AI has significantly enhanced the efficiency and quality of forensic appraisal, demonstrating immense potential, it simultaneously encounters substantial dilemmas at both the technical and legal levels. Technically, algorithmic bias and the “black box” phenomenon undermine the reliability and interpretability of expert opinions, accompanied by inherent data privacy risks. Legally, lagging regulatory frameworks result in ambiguous subjective status and liability attribution for AI-assisted appraisals, while existing procedural rules and cross-examination mechanisms fall short of effectively addressing such opinions. To this end, targeted institutional improvements should be advanced across four dimensions: regulating algorithmic data, mandating the disclosure of algorithms and data utilized in appraisals, refining the rules of cross-examination in trials, and constructing a sound mechanism for liability allocation. These measures will ultimately foster the benign development of AI-assisted forensic appraisal and safeguard judicial justice.