Based on "meaningful human control", which is the core concept of artificial intelligence ethics, the ethical difficulties in the algorithm design stage for autonomous driving are summarized in the article. The feasibility of "meaningful human control" applied to autonomous vehicles is analyzed. On this basis, using the two conditions of "tracking" and "tracing", a "meaningful human control" framework is constructed around "accountability and transparency" and "value-sensitive design", which may provide systematic methodology guidance for algorithm design of autonomous vehicles.

autonomous vehicles  /  ethical decision-making  /  trolley problem  /  meaningful human control  /  value-sensitive design