摘要
随着生成式AI技术的持续进步,人文研究的数字人文范式正在向“AI人文”转型。传统的数字人文通过技术手段,成功将人文文献转化为数字化和数据化的形式,实现了对文献的高效处理与利用。AI人文不仅是数字人文的升级版,更有望实现AI4R(AI for research),推进人文信息分析的自动化和智能化,具有知识生产的潜能。生成式AI实现了人类经验的向量化,通过模式识别、信息重组、预测推断和语言生成技术等,具有了辅助参与知识体系建构、提供趋势预测和决策参考的重要功能。不过,目前AI人文的基本方式还是人机交互辅助生成,仍是“需要人的AI”,人依然处于中心性、支配性和主导性地位。同时,“拥有AI的人文”也亟待应对算法黑箱、算法偏见、AI幻觉,以及可能的AI自我意识的觉醒等问题。
With the continuous progress of generative AI technology,the digital humanities paradigm of humanities research is transforming into“AI humanities”.Traditional digital humanities have successfully transformed humanistic literature into digital and data-driven form by technical means,thus realizing efficient processing and utilization of literature and information.AI humanities is not only an upgraded version of digital humanities,but also is expected to achieve“AI4R”(AI for Research),which promotes the automation and intelligence of humanities information analysis and possesses potentials of knowledge production.Generative AI has achieved vectorization of human experience.By pattern recognition,information reorganization,prediction and inference as well as language generation technology,it has important functions of assisting in the construction of knowledge system and providing reference for trend prediction and decision making.However,the basic way of AI humanities lies in human-computer interaction assisted generation,which is still the“AI that needs man”,and“man”is still in the central,leading and dominant position.At the same time,the“humanities with AI”also needs to deal with problems such as algorithmic black boxes,algorithmic bias,AI illusion and the possible awakening of AI self-awareness.
出处
《东南学术》
北大核心
2024年第4期42-51,246,共11页
Southeast Academic Research
基金
国家社会科学基金重点项目“当代中国对话主义文学理论的话语建构研究”(项目编号:22AZW003)
上海市教育委员会科研创新计划人文社会科学重大项目“新媒体艺术理论基本问题研究”(项目编号:2023SKZD15)。
