论生成式人工智能应用的国家安全审查

On National Security Review System of Generative Artificial Intelligence Applications

生成式人工智能已进入应用层面,给人们带来诸多便利的同时,亦带来诸多挑战,国家安全风险即其中一项。生成式人工智能在政治安全、数据安全、经济安全、文化安全等方面均可能产生风险。防范和化解国家安全风险需要系统推进,其中国家安全审查制度因其个案性特征而可以发挥重要作用。现行国家安全审查制度体系包括外商投资安全审查、网络安全审查、数据安全审查、国外上市安全审查、兜底性国家安全审查等,在一定情形下可适用于生成式人工智能应用。但现行制度适用时面临覆盖不全、无法作为常规手段、规则不清晰可能导致阻碍发展等问题,为维护国家安全,平衡好防范和化解国家安全风险与促进发展的关系,应当在时机成熟时建立更为清晰的针对生成式人工智能应用的国家安全审查规则。

Generative artificial intelligence(GAI)has entered the application stage,bringing many conveniences to people and also posing numerous challenges,including risks to national security.GAI applications may cause risks in political security,data security,economic security,cultural security and other aspects.Preventing and mitigating national security risks requires systematic efforts,where the national security review system can play an important role due to its case-by-case approach.The current national security review system includes foreign investment security review,network security review,data security review,foreign listing security review,comprehensive national security review,etc.,which can be applied to GAI applications under certain circumstances.However,the existing system faces issues such as incomplete coverage,inability to be a routine measure,and unclear rules that may hinder development.To safeguard national security and strike a balance between preventing and mitigating national security risks and promoting development,clearer national security review rules specifically tailored for GAI applications should be established at the opportune moment.