想象的困境:生成式人工智能世代的价值教育

The Imagined Dilemma:Value Education for the Generation of Generative Artificial Intelligence

“道德物化”一直是人工智能的重要取向。以ChatGPT为代表的新一代生成式人工智能也正试图通过牺牲一定通用性能的“对齐税”的策略,来表达人类道德伦理和情感价值。然而,代码系统的横亘,使学生从人工智能所获得的知识技能只能是“二手的非经验”,从而造成学生与生活世界的“短路”,直接面临着“去道德化”的风险。“对齐税”策略似乎并不能够还原与拯救人类价值。知—情—意本当是一个整全。“知识归AI,智慧归人类”的策略一开始就落入了二元分裂的窠臼,而缺乏知识必将扭曲人们构建世界的心智地图,道德伦理和情感价值教育因此可能只是一厢情愿。生成式人工智能世代更应该朝向生活世界,将“人之为人”植根于整体世界的语境中,促进学生的全面而自由的发展。本研究并非作出一系列判断,毋宁是对生成式人工智能世代的价值教育提出的一连串疑问。

Moral objectification has always been an important orientation of artificial intelligence.The new generation of generative artificial intelligence,represented by ChatGPT,is also attempting to express human moral ethics and emotional values by sacrificing certain universality of the alignment tax strategy.However,the existence of code systems leads to the knowledge and skills that students acquire from AI being nothing more than second-hand non-experience,resulting in a disconnect between students and the life world and directly facing the risk of demoralization.The alignment tax strategy does not seem to be able to restore and save human values.As a whole,knowledge,emotions,and intentions should be integrated,yet the strategy of“knowledge belongs to AI,wisdom belongs to humans” falls into the trap of binary division from the beginning,and a lack of knowledge will inevitably distort people's cognitive maps for constructing the world.Therefore,value education in moral ethics and emotional values may be just wishful thinking.The generation of generative artificial intelligence should move towards the life world,anchoring“the essence of being human”in the context of the overall world and promoting students' comprehensive and independent development.This research does not provide a series of judgments but rather raises a series of questions about the value education of the generation of generative artificial intelligence.