中国电化教育

ChatGPT的成功引发了教育界对生成式人工智能广泛而又热烈的讨论。在此背景下，以深度学习为线索对生成式人工智能何以构成教育奇点予以追问，意在从前提上把握生成式人工智能之于教育的突破和局限。从深度学习的突破来看，生成式人工智能之于教育的深层价值在于擘画了指向算法决策的新方向、指向大模型的新范式、指向智能机器的新赛道；从深度学习的局限来看，教育极有可能因为生成式人工智能而面临不可知、不可用和不可信的风险。这对热衷求证生成式人工智能何以构成奇点的教育而言，仍不足以构成可谓确切的答案；但对试图诉诸更有深度的概念、问题、目的以创造奇点的教育而言，却不失为有益的启示。归根究底，与作为时下人工智能重要方向之一的深度学习相比，对人工智能跨领域、长时间、多层次、辩证性的深度理解更有可能引申出教育奇点。

The success of ChatGPT has sparked widespread and heated discussions in the education community about generative artificial intelligence. In this context, using deep learning as a clue to question how generative artificial intelligence constitutes the singularity of education, the aim is to grasp the breakthrough and limitations of generative artificial intelligence in education from the previous context. From the perspective of breakthroughs in deep learning, the deep value of generative artificial intelligence in education lies in charting new directions towards algorithmic decision-making, new paradigms towards large models, and new tracks towards intelligent machines; From the limitations of deep learning, education is highly likely to face unknown, unavailable, and untrustworthy risks due to generative artificial intelligence.For educators who are enthusiastic about verifying how generative artificial intelligence constitutes a singularity, this is still insufficient to provide a definitive answer; But for education that attempts to appeal to deeper concepts, problems, and purposes to create singularities, it can be a useful inspiration. Ultimately, compared to deep learning, which is one of the important directions of artificial intelligence today, a deep understanding of AI's cross disciplinary, long-term, multi-level, and dialectical nature is more likely to lead to educational singularities.