ChatGPT赋能智慧法院建设:机遇、挑战和规制

ChatGPT-enabled smart court construction:Opportunities, challenges, and regulations

ChatGPT类生成式人工智能在海量的数据、优质的算法与强大的算力支撑下能够实现多项技术赋能。但ChatGPT类生成式人工智能是否能够适用到智慧司法领域,以及在何种程度上将其嵌入,尤其是该项技术在赋能智慧法院建设过程中的边界如何,这对于深层次建设智慧法院具有重要的前提性意义。ChatGPT类生成式人工智能技术通过审判活动的智能化运用,可促进可视正义的实现;通过诉讼服务的可及性,可实现司法为民理念;通过公平与效率的动态平衡,可实现能动司法目标。然而,ChatGPT类生成式人工智能在带来上述诸多技术红利的同时,也引发了一系列新的风险,具体可以划分为由数据“幻觉性”引发的一般风险、由技术垄断性导致的重要风险以及由数据不可控产生的核心风险。对这些风险,应根据不同的风险特征采取相应的规制路径:可以通过对训练数据库的标准化构建以实现数据的“幻觉性”规制,通过辅助性原则的引入而实现技术规制,通过对服务提供者科加义务以实现对数据不可控的数据规制。即运用数据“幻觉性”规制、技术规制与数据规制,实现对ChatGPT类生成式人工智能技术嵌入的一般风险、重要风险与核心风险的防范,以期完成ChatGPT类生成式人工智能技术嵌入智慧法庭建设的风险最小化与技术赋能最大化,最终实现该项技术高质量赋能智慧法院建设。

ChatGPT-like generative AI can achieve multiple technical empowerments with the support of massive data,high-quality algorithms,and powerful computing power.However,whether ChatGPT-like generative AI can be applied to the field of smart justice and to what extent it can be embedded,especially the boundaries of this technology in empowering the construction of smart courts,are important prerequisites for the deep construction of smart courts.ChatGPT-like generative AI technology can promote the realization of visual justice through the intelligent application of trial activities;realize the concept of justice for the people through the accessibility of litigation services;and achieve the goal of active justice through the dynamic balance of fairness and efficiency.However,while ChatGPT-like generative AI brings many technological dividends,it also raises a series of new risks,which can be specifically divided into general risks caused by the illusion of data,important risks caused by technological monopoly,and core risks caused by uncontrollable data.For these risks,corresponding regulatory paths should be adopted according to different levels of risk:data illusion regulation can be achieved through the standardized construction of training databases,technical regulation can be achieved through the injection of auxiliary principles,and data regulation can be achieved by imposing obligations on service providers to achieve uncontrollable data.That is,using data illusion regulation,technical regulation,and data regulation to prevent general risks,important risks,and core risks in the embedding of ChatGPT-like generative AI technology,with a view to minimizing the risks and maximizing the technical empowerment in the construction of smart courts,and ultimately achieving high-quality empowerment of smart courts with this technology.