我国人工智能安全检测能力现状及治理建议

Current Status and Governance Suggestions for Artificial Intelligence Security Detection Capability in China

人工智能大模型现象级的火爆,促使人工智能与实体经济的加速融合与应用。人工智能安全治理攸关全人类命运,通过测试评估防范安全风险,推进人工智能发展与提升人工智能安全治理能力已成为全人类的共识。针对人工智能安全威胁由局部攻击向系统化协同攻击演化,导致单一的检测与防护技术无法应对复合攻击的问题,因此加速提升人工智能安全检测能力,保障人工智能安全刻不容缓。通过梳理人工智能安全检测政策背景、安全风险、能力现状,提出了治理建议,对人工智能产业安全发展具有重要意义。

The phenomenon level popularity of large-scale AI(Artificial Intelligence)models accelerates the integration and application of AI with the real economy.AI security governance is crucial to the fate of all humanity,and it becomes a global consensus to prevent security risks through testing and evaluation,and to promote the development of AI and enhance its security governance capabilities.AI security threats evolve from local attacks to systematic and collaborative attacks,resulting in a single detection and protection technology that cannot cope with the problem of composite attacks.In order to address this problem,it is urgent to accelerate the improvement of AI security detection capabilities and ensure AI security.It’s of great research significance and practical application value for the secure development of AI industry to review the policy background,security risks,current status of capabilities,and governance suggestions for AI safety detection.