生成式人工智能的国家安全风险及其对策

National Security Risks of Generative Artificial Intelligence and Countermeasures

以ChatGPT为代表的生成式人工智能成为专用人工智能迈向通用人工智能的关键,拉开了强人工智能时代的序幕。生成式人工智能开创了“大模型”主导的内容生成时代,给政治安全、军事安全、网络安全、经济安全、社会安全等国家安全领域带来了诸多风险与挑战。为防控风险,应遵循人本逻辑制定防控生成式人工智能风险的“一体化”顶层战略,遵循法治逻辑健全防控风险的“全方位”制度体系,遵循共治逻辑构建防控风险的“同心圆”治理结构,遵循科技逻辑强化防控风险的“全周期”过程监管。

The generative artificial intelligence represented by ChatGPT has become the key to moving from dedicated artificial intelligence to general artificial intelligence,marking the beginning of the era of strong artificial intelligence.Generative AI ushered in an era of"big model"dominated content generation,and brings many risks and challenges to national security fields such as political security,military security network security,economic security,and social security For preventing and controlling generative AI risks,it is necessary to follow the logic of people-oriented to develop an"integrated"top-level strategy,follow the logic of the rule of law to improve the"comprehensive"system for preventing and controlling risks,follow the logic of co-governance logic to build a"concentric circle"governance structure for preventing and controlling risks,and follow the scientific logic to strengthen the"full cycle"process supervision for preventing and controlling risks.