人工智能在军备控制领域的争议与挑战

随着人工智能技术的发展，当前国际社会对管控以“杀手机器人”为代表的致命性自主武器系统的呼声日益强烈，对于这类武器系统可能在未来引发伦理和安全危机的担忧逐渐增多。针对致命性自主武器，最严格的观点是全面禁止其研发和使用，以防止“杀手机器人”的出现；较宽松的观点是限制其使用、转让、扩散和购买；介于两者之间的观点是暂停其研发或使用，待相关管控规则形成后，再开展相关活动。本文认为，目前，国际社会就这一军控问题达成共识依然任重道远。根据现实情况，我国不宜过早寻求限制相关技术的研发工作，而应力求先确立管控致命性自主武器部署、使用、流通等方面的规则。

Controversy and challenge of AI in the field of arms control

With the development of AI technology, there are mounting calls in the current international community for the control of Lethal Autonomous Weapons (LAWs), represented by Lethal Autonomous Robots (LARs). Meanwhile, the concern that LARs may lead to future ethic and security crisis has gradually increased. The most stringent control proposal for LARs is an overall prohibition of their research, manufacture, deployment and usage to prevent them from emergence; a relaxed control approach aims to restrict their use, transfer, diffusion and purchase activities; and a way in between is a temporary suspension of their research, manufacture and usage until relevant control rules are forward. At present, there is still a long way to reach consensus on the control of LARs at the international level. According to the actual situation, it is inappropriate to seek to restrict the research and development of related technologies too early. Instead, it is necessary to establish rules governing the deployment, use and circulation of LAWs.