算法推荐引发道德认知偏差的症候、成因与规制

Symptoms,Causes,and Regulations of Moral Cognitive Bias Caused by Algorithm Recommendations

认知主体的技术具身与分化、认知对象的数字解蔽与遮蔽、认知环境的精准架构与固化、认知载体的价值中立与偏向是算法推荐技术赋予道德认知的双重语境,不仅为精准把握人的道德认知需求、驱动人的道德提升提供了现实可能,还引发了道德认知的偏差问题。算法推荐技术引发的道德认知偏差,以话语陷阱、情感极化、信息茧房为表征,以认知图式异化、认知理性缺位、认知空间窄化为生成逻辑,其生成与算法剥夺造成智能依赖、技术僭越带来主体式微、资本外溢引发价值遮蔽密切相关。规制算法推荐引发的道德认知偏差,需要以算法素养提升破解算法宰制、以人的逻辑超越算法遮蔽、以核心价值观引领算法方向,重塑人的主体性及其与算法技术、算法平台的互动关系。

The technological embodiment and differentiation of epistemic agencies,the digital revelation and concealment of cognitive objects,the precise architecture and solidification of cognitive environments,and the value neutrality and bias of cognitive carriers are the dual contexts that algorithmic recommendation technology endows with moral cognition.This not only provides a realistic possibility for accurately grasping people's moral cognitive needs and driving their moral improvement,but also triggers the problem of deviation in moral cognition.The moral cognitive bias caused by algorithmic recommendation technology is characterized by the discourse trap,emotional polarization,and information cocoon,and take cognitive schema alienation,cognitive rationality deficiency,and cognitive space narrowing as generative logic.Its generation is closely related to the dependence of artificial intelligence caused by algorithmic deprivation,the decline of the subject caused by the arrogation of technology,and obscuration of value caused by capital spillover.The moral cognitive bias caused by regulatory algorithm recommendations requires the improvement of algorithm literacy to break through algorithm domination,to surpass algorithm concealment with human logic,to lead algorithm direction with socialist core values,and to reshape human subjectivity and its interactive relationship with algorithm technology and algorithm platforms.