民航人工智能的潜在风险与应对策略——以波音MCAS为例

Potential Risks and Countermeasures of Artificial Intelligence in Civil Aviation——Take Boeing MCAS as an Example

2018年10月至2019年3月短短半年之内发生了两起极其相似的空难,事故的原因均指向了波音公司的人工智能系统MCAS,这让人们意识到民航人工智能存在一系列潜在风险。通过对MCAS引起的两起空难的分析,可以发现民航人工智能可能会存在设计缺陷、科技伦理风险、智能权限不当、监管失职、飞行员应对失策等方面的问题。为了应对上述潜在风险,首先要设计负责任的民航人工智能,其次要在技术和制度两个层面加强对民航人工智能的监管,最后要建立智能飞行人机规范,保证人掌握最高权限,使人工智能始终处于人的监督之下。

Within just six months,from October 2018 to March 2019,two very similar air crashes oc‐curred,the causes of the accidents all point to Boeing’s artificial intelligence system MCAS,which makes people realize that there are a series of potential risks in civil aviation artificial intelligence.Through the analysis of the two crashes caused by MCAS,some problems due to civil aviation artificial intelligence may be found,such as design defect,ethical risks of science and technology,improper intelligent authority,regu‐latory dereliction of duty,and pilot’s improper countermeasures.In order to cope with the above potential risks,we must first design a responsible civil aviation artificial intelligence.Secondly,we must strengthen the supervision of civil aviation artificial intelligence at the technical and institutional levels.Finally,we must establish intelligent flight man-machine norms to ensure that people have the highest authority and that artificial intelligence is always under human supervision.