欧美大学面对ChatGPT的挑战与抉择-清华大学教育研究

文章信息/Info

Title:: Challenges and Choices Faced by European and American Universities from ChatGPT

Author(s):: WANG Rong¹ LI Lian-ming²; 1.Chengxian School, Southeast University;2.Office of Nanjing University

摘要:: 世界范围内以ChatGPT为代表的人工智能技术发展如火如荼、方兴未艾。欧洲多国在最初的保守氛围中,以牛津大学和剑桥大学为代表,在明令禁止使用ChatGPT之后不久,便在新技术革命的澎湃浪潮中及时发布罗素大学集团AI使用指南。美国以哈佛大学和普林斯顿大学为代表,前者经过短暂禁止之后,迅速在计算机课程中设计出ChatGPT“橡皮鸭子”助教并开始采用ChatGPT教育工具;后者则发明出针对ChatGPT用于作弊的ChatZero识别器,检测准确率达到98.6%。这些大学及其相关学者的心路历程、应对策略与研究成果,再加上斯坦福大学这一AI“世界冠军”的成功经验、OpenAI最新款GPT-4o的新功能添加,以及国会议员竭力推动联邦政府创建AI领域“曼哈顿工程”等,都会对大国硬实力的蓄积产生不容小觑的影响。随着我国开源性DeepSeek横空出世,ChatGPT所面临的竞争与压力正与日俱增。

Abstract:: The global development of AI technology, represented by ChatGPT, has been thriving. Initially adopting a conservative stance, several European countries—represented by the University of Oxford and the University of Cambridge—issued bans on ChatGPT, but soon responded to the rising wave of technological revolution by releasing AI usage guidelines through the Russell Group of Universities. In the United States, Harvard University briefly banned ChatGPT before designing a “rubber duck” TA based on it for use in computer science courses. Meanwhile, Princeton University developed “GPTZero,” a tool aimed at detecting ChatGPT-based cheating, with a reported accuracy rate of 98.6%. These universities’ experiences—including their strategies, reflections, and research achievements—combined with Stanford University’s leadership in AI, the newly released capabilities of OpenAI’s GPT-4o, and the U.S. Congress’s push for a “Manhattan Project” in AI, are contributing significantly to the buildup of hard power in major countries such as the United States and China. Meanwhile, ChatGPT has been facing fiercer pressure of competition since DeepSeek roared across the horizon.