Research on the Criminal Imputation of Crimes Involving Level 3 Autonomous Vehicles

Engineering notes

Key topics: autonomous driving, autonomous vehicle. See the paper for implementation details and experimental results.

Chinese explanation / 中文解读

中文解读待补充：本站会优先为端到端自动驾驶、BEV感知、3D目标检测、轨迹预测、路径规划、LiDAR感知等高价值论文补充中文说明。

Original abstract

The technical characteristics of Level 3 autonomous driving, namely "human-machine co-driving and dynamic takeover", have subverted the logic of criminal imputation in traditional traffic accident cases, and relevant accidents have highlighted core dilemmas in the application of criminal law. Specifically, at the level of liable subjects, current norms fail to clarify the standards of duty of care for drivers, automobile manufacturers and algorithm designers; algorithmic black boxes and data barriers cause difficulties in evidence production and responsibility shifting, and the qualification of artificial intelligence as a criminal subject remains controversial. At the level of subjective imputation, algorithmic black boxes invalidate the standard of concrete foreseeability of consequences under the traditional old fault theory, hindering the presumption of subjective culpability. At the level of causality, algorithmic black boxes block the attribution chain, and probabilistic decision-making breaks through traditional rules of causality determination. The criminal punishability of artificial intelligence should be clearly denied, the proportion of subject liability should be defined according to different scenarios, the standards of subjective imputation should be reconstructed with the new fault theory and the duty of care should be refined. Impediments to imputation should be resolved through rules such as the pre-duty of algorithm interpretability and the inversion of the burden of proof, so as to provide a feasible application approach for such cases.

Links and sources

Comments