Kill chains constitute a good model for analyzing the effectiveness of combat elements such as weapons and systems. Owing to the lack of a uniform definition of the concept of intelligent weapons, this study develops a method for judging intelligent weapons based on the kill-chain framework. In view of the arms control controversies caused by the difficulty in identifying intelligent weapons, two risk factors that could lead to loss of control over intelligent weapons were analyzed: the loss of control over autonomous weapons, and the existence of uncontrollable factors and defects in autonomous weapons. The accountability dilemma of intelligent weapons, that is, the allocation of responsibility among commanders, programmers, and intelligent weapons, was also analyzed. Based on the above, the modes by which intelligent weapons could be deployed in future confrontations were addressed. Proposals for risk management and arms control with respect to intelligent weapons were also provided. The “weapon + AI” and “AI + weapon” deployment modes were examined, and the suggestions for risk management included limiting the scope of intelligent weapons, embedding physical restriction systems in intelligent weapons, limiting training data for intelligent weapons, and controlling the defense acquisition process.
郝英好.基于打击链的智能武器风险分析与控制方法研究[J].国防科技,2021,42(1):37-42；HAO Yinghao. Kill-chain-based analysis of the risks of intelligent weapons and methods for their control[J]. National Defense Technology,2021,42(1):37-42复制