人工智能“类人化”的人学反思

Anthropology Reflection on Anthropomorphic Artificial Intelligence

检视人工智能类人化问题,必须首先回答“人是什么”的问题。以马克思的人学视角来看,人是兼具自然性、社会性、道德性、审美性和自我超越性等多重规定的综合性存在。虽然人的技术理性能力赋予了人工智能以类人化契机,但也正是人工智能单向度的理性能力,决定了它不仅在本质上与人存在着较大的异质性,而且暴露出其类人化的程度和限度。面对前途未定的人工智能发展前景,若要对人工智能类人化发展向度和限度做出合理的科学评估和价值判断,我们只有坚持价值理性和工具理性相统一,从马克思人学立场反思“人应何为”,才能保证人工智能类人化不仅能够“向人而生”,而且始终“为人而生”。

From the perspective of humanity defined by Marx,a human is a comprehensive existence with multiple stipulations of nature,sociability,morality,aesthetics,and self-transcendence.Although the technically rational nature of humans endows the opportunity of anthropomorphic artificial intelligence(Al),the anthropomorphic AI not only has a great heterogeneity from humans in essence but also exposes the limitations of humanization because of the AI's unidirectional rational capacity.Facing the uncertain prospect of Al in the future,if we want to make reasonable scientific evaluations and value judgments regarding the direction and limitation of the anthropomorphic Al,it is required to adhere to a unification of both value rationality and instrumental rationality and reflect"what a human should be?"From the perspective of Marx's anthropology so that AI can not only be"similar to humans"but also always"live for humans".