新工科背景下生成式AI融入“工具–协作–批判”教学与科研协同育人实践——以汽车工程材料课程为例

Integrating Generative AI into “Tool-Collaboration-Critique” Teaching and Research Practice under the New Engineering Background—Taking Automotive Engineering Materials Course as an Example

摘要: 新工科建设对材料学科人才培养提出了更高要求，然而当前汽车工程材料课程教学中，生成式AI的应用普遍停留在“智能问答”式的浅层辅助，与真实科研实践严重脱节。为此，本文提出并实践了一种从“智能问答”走向“实验求证”的研究性教学模式。该模式以“AI辅助生成可验证假设–团队协作优化–实验实证批判修正”为核心链路，将真实科研任务——“AI辅助设计新型锂电池电解质材料”——贯穿教学全程。实践表明，学生借助生成式AI工具成功筛选出三种新型化合物，经实验室合成与电化学性能测试，实测数据显示其性能较传统材料显著提升，相关教学案例成果已整理并拟投稿至国际EI会议进行学术交流。该模式有效弥合了AI应用与科研训练之间的鸿沟，实现了“以研促学、以学助研”的良性循环，为新工科背景下材料类课程的教学改革提供了可复制的研究性教学路径。

Abstract: The development of New Engineering Education imposes increasingly rigorous demands on talent cultivation in materials science. Nevertheless, within the current teaching framework of the Automotive Engineering Materials course, the application of generative artificial intelligence remains predominantly confined to a superficial assistive level—typically in the form of intelligent question‑answering—which is markedly disconnected from authentic research practice. In response, this study presents and implements a research‑oriented pedagogical model that transitions from “intelligent Q&A” to “experimental validation.” The model is structured around a core loop: AI‑assisted generation of verifiable hypotheses, team‑based collaborative optimization, and empirical verification with critical revision. A real‑world research task—the AI‑assisted design of novel electrolyte materials for lithium batteries—is seamlessly integrated throughout the entire instructional process. Empirical results show that students, using generative AI tools, successfully screen three novel compounds. Subsequent laboratory synthesis and electrochemical performance testing reveal that the measured properties are substantially improved over those of conventional materials. The outcomes of this pedagogical case are systematized and are intended for submission to an international EI‑indexed conference for scholarly dissemination. This model effectively narrows the gap between AI application and research training, thereby achieving a virtuous cycle of “research‑enhanced learning and learning‑supported research.” It offers a replicable pathway for research‑oriented teaching reform in materials‑related curricula within the context of New Engineering Education.