网络暴力犯罪中平台算法推荐行为的刑法归责

Criminal Imputation of Platform Algorithmic Recommendation Behavior in Cyber Violence Crimes

摘要: 现今社会，有必要拓展网络平台间接刑事归责的教义学路径，威慑平台及时采取防范网暴风险的技术措施。针对平台“明知”网暴信息的存在，仍为其传播提供帮助或放任其传播的行为，应当以片面帮助犯追究其刑事责任。有必要对网络平台算法推荐行为的归责依据与归责困境进行研究，分析其在客观追责与主观故意方面认定的困境，并探讨相应的解决办法。据此，提出应在客观层面采用“相当因果关系说”并通过“阈值”设定与技术手段量化算法推荐的行为贡献度；在主观层面明确“明知”的认定标准，合理运用推定破解算法“黑箱”带来的证明困境，明确平台间接刑事责任的认定标准，并通过责任限缩合理地划分应对网络平台刑事归责的情形。

Abstract: In contemporary society, it is necessary to activate the model of indirect criminal imputation for online platforms, thereby deterring platforms and compelling them to promptly adopt technical measures to prevent the risk of cyber violence. Where a platform, clearly knowing of the existence of cyber violence information, still provides assistance for its dissemination or allows such information to spread, criminal liability should be pursued on the basis of unilateral aiding. It is essential to examine both the grounds for imputing liability for algorithmic recommendation behaviors of online platforms and the dilemmas involved in such imputation, analyzing the difficulties in establishing objective imputation and subjective intent, and exploring corresponding solutions. Accordingly, it is proposed that, at the objective level, the “theory of adequate causation” should be adopted, and the degree of behavioral contribution of algorithmic recommendations should be quantified through the setting of “thresholds” and technical means; at the subjective level, the criteria for determining “knowledge” should be clarified, presumptions should be reasonably employed to overcome the proof dilemma arising from the algorithmic “black box”, the indirect criminal liability of platforms should be tightened, and the circumstances triggering the criminal imputation of online platforms should be reasonably delineated through the limitation of liability.