The integration of artificial intelligence (AI) into educational environments has initiated a shift in how students acquire knowledge, moving from active inquiry to the passive reception of algorithmic outputs.

將人工智慧（AI）整合至教育環境中，已導致學生獲取知識的方式發生轉移，從主動探究轉向被動接收演算法的輸出結果。

Historical pedagogical models emphasized the necessity of intellectual friction, wherein the acquisition of knowledge was contingent upon rigorous effort, source verification, and the navigation of uncertainty. The current transition toward AI-mediated learning has introduced a 'silent curriculum' characterized by the removal of these cognitive obstacles. This phenomenon is evidenced by data from Oxford University Press, indicating that while 90% of surveyed teenagers perceive academic skill development via AI, 60% acknowledge a concomitant decline in other competencies. In the Indian context, the 2025 Bharat Survey for EdTech and the FICCI-EY AI Adoption Survey report high utilization rates—86% among students—contrasted with a significant deficit in faculty AI literacy, with only 17% of educators identifying as advanced users.

傳統的教學模式強調「智力摩擦」的必要性，即獲取知識必須依賴於艱苦的努力、來源驗證以及對不確定性的處理。目前的 AI 介導學習轉型引入了一套「隱形課程」，其特點在於消除了這些認知障礙。牛津大學出版社的數據證明了這一現象：雖然 90% 的受訪青少年認為透過 AI 能發展學術技能，但 60% 承認其他能力隨之下降。在印度的情境中，2025 年 Bharat EdTech 調查與 FICCI-EY AI 採納調查報告指出，學生端的使用率極高（達 86%），但與之相對的是教職員 AI 素養嚴重缺乏，僅有 17% 的教育工作者認同自己是高級使用者。

Academic discourse has identified several critical cognitive risks associated with this transition. Researchers at Wharton have termed the tendency to accept AI outputs without critical evaluation as 'cognitive surrender,' supported by experimental data showing a 79.8% acceptance rate of AI recommendations even when intentionally erroneous. Furthermore, Lucy Gill-Simmen of Royal Holloway, University of London, posits that the primary risk is 'epistemic atrophy,' wherein the habitual processes of knowledge acquisition weaken. This leads to an 'illusion of understanding,' where students produce fluent responses but lack the capacity to apply reasoning in novel contexts. The OECD has similarly highlighted the risk of 'cognitive offloading,' suggesting that children may delegate tasks before the foundational mental models required to perform those tasks have been established.

學術論述已指出此次轉型相關的幾項關鍵認知風險。華頓商學院的研究人員將這種不經批判性評估即接受 AI 輸出的傾向稱為「認知投降」，實驗數據顯示，即使 AI 的建議被刻意設錯，接受率仍高達 79.8%。此外，倫敦大學皇家霍洛威學院的 Lucy Gill-Simmen 主張，主要風險在於「認識萎縮」，即習慣性的知識獲取過程逐漸弱化。這導致了「理解錯覺」，使學生能產出流暢的回答，卻缺乏在創新情境中應用推理的能力。OECD 同樣強調了「認知卸載」的風險，認為孩童可能會在建立執行該任務所需的基礎心理模型之前，就將任務委託給 AI。

Institutional responses, particularly within India, have focused on the integration of AI and Computational Thinking into primary and secondary curricula. However, experts suggest that true AI literacy must transcend operational proficiency to include the ability to identify algorithmic hallucinations and systemic biases. The proposed strategic rapprochement involves transitioning AI from a replacement for thought to a 'co-intelligence' tool. This framework necessitates an intentional curriculum that prioritizes real-world experiences, collaborative problem-solving, and the critical interrogation of AI outputs to ensure that cognitive autonomy is maintained.

機構的回應（特別是在印度）集中於將 AI 與計算思維整合至中小學課程中。然而，專家建議，真正的 AI 素養必須超越操作熟練度，應包括識別演算法幻覺與系統性偏見的能力。擬定的策略調整方向是將 AI 從「思考的替代品」轉向「協作智能」工具。此框架需要一套有意識的課程，優先考慮現實世界的經驗、協作解決問題，以及對 AI 輸出進行批判性質詢，以確保認知自主權得以維持。

The current educational landscape faces a critical tension between the efficiency of AI-generated answers and the preservation of deep, independent critical thinking skills.

目前的教育景觀正面臨著一個關鍵矛盾：AI 生成答案的高效率與保留深層、獨立批判性思考能力之間的拉鋸。