人工智能安全治理挑战与对策

Challenges and Countermeasures of Artificial Intelligence Security Governance

人工智能发展几经起伏,近年来再一次引发学术界和产业界的高度关注,其技术正在快速地应用于各个领域,已成为各国实现产业转型升级的新一轮战略性技术.但是以机器学习为核心技术的人工智能深度应用所引发的技术风险和社会风险日益凸显.从人工智能技术潜在安全漏洞、过度滥用和社会伦理3个方面,综述分析人工智能所面临的安全风险及其治理现状.进一步针对人工智能安全治理问题,从技术、标准和法律几个维度提出解决方案及建议,旨在为人工智能安全治理体系的建立及行业应用提供思路,也为人工智能安全技术研究提供探索方向.

The development of artificial intelligence has gone through several ups and downs.In recent years,it has once again attracted the great attention of academia and industry.Its technology is being rapidly applied in various fields and has become a new round of strategic technology for countries to realize industrial transformation and upgrading.However,the indepth application of artificial intelligence with machine learning as the core technology has brought about increasingly prominent technical and social risks.This paper summarizes and analyzes the security risks faced by artificial intelligence and its governance status from three aspects:potential security vulnerabilities,excessive abuse,and social ethics.To further deal with the issue of AI security governance,this paper puts forward solutions and suggestions from the perspectives of technology,standards,and laws,aiming to provide an idea for the establishment of AI security governance systems and industrial applications.Meanwhile,this paper also gives a direction for the exploration of AI security technology research.