可信人工智能标准体系建设研究

Research on the Construction of Trusted AI Standards System

人工智能的发展和进一步应用,已经成为不可逆的趋势。人工智能带来福利的同时,也带来一定的风险。通过可信人工智能标准体系构建的方法,可降低人工智能所带来的风险。可信人工智能要求人工智能系统具备可解释性、无歧视性和数据安全性。域外可信人工智能标准体系建设主要存在伦理概念标准建设和技术标准建设两种途径,并已经有相关的可信人工智能标准体系建设总纲。我国可以参照域外的经验,建设可信人工智能标准体系总纲,加强人工智能伦理标准体系建设和技术标准体系建设,确保人工智能系统的可信性。

The development and further application of artificial intelligence(AI)has become an irreversible trend.While AI brings benefits,it also brings certain risks.The risk brought by AI can be reduced through the method of building a trusted AI standards system.Trusted AI requires that artificial intelligence systems have explainability,non-discrimination,and data security.There are mainly two ways to build an extraterritorial trusted AI standards system:the construction of ethical concept standards and the construction of technical standards,and there is already a relevant general outline for the construction of a trusted AI standards system.China can refer to the overseas experience to build a general outline of a trusted AI standards system,strengthen the construction of an AI ethical standards system and a technical standards system,and ensure the credibility of AI systems.