【摘要】 经历了半个世纪发展历程的教育人工智能在21世纪更加备受瞩目，而以Chat GPT为代表的生成式人工智能则把这种炒作推向新高潮。“教育离不开人工智能”“教育要拥抱人工智能”“人工智能必将颠覆教育”……诸如此类的预言似乎已经成为学界主流话语。本文在回顾教育人工智能历史、区分人工智能与人类智能和辨识教育的“小世界”和“大世界”观等基础上，对教育人工智能十大预言/悖论进行批判分析，并反思我们需要什么样的教育人工智能观。更多还原

【Abstract】 Artificial intelligence in education(AIED) which can be traced back to over 50 years ago has increased its momentum since the turn of the century. Nevertheless, it is the release of ChatGPT by OpenAI in November, 2022 that has pushed the hype around AIED to its crest, putting it in the spotlight of the wider global society. AI is essential to education,education should embrace AI, and AI is to disrupt education. Predictions of the kind constitute the mainstream discourse in academia. This article first reviews the history of AIED, distinguishes between AI and human intelligence, and examines the nature of education from the perspective of Leonard Savage’s “small worlds” and “big worlds”. It then goes on to critique ten popular predictions/paradoxes of AIED, namely 1) Education can be automated; 2) Data and algorithm are objective and neutral; 3) AI systems are not impartial and neither are human educators; 4) AI is good at personalizing education; 5) AI renders some conventional skills and assessment tools redundant; 6) AI is time-saving; 7) AI can reduce cost; 8) AI enables equity in education; 9) AI applications are to be normalized in education; and 10) AI will inevitably transform the future of education – for the better. The article concludes with the author’s insights into AIED.更多还原