面向能力训练的人机协同作业设计与实证研究

Design and Empirical Research on Human-Machine Collaborative Work for Ability Training

摘要: 针对传统工科作业难以支撑复杂工程问题解决能力培养的现状，本研究探索生成式人工智能赋能作业设计的路径。以协同论为理论基础，明确人机主次，非线性交互的设计思想；以脚手架理论指导作业任务梯度与人机交互反馈的递进设计，构建了涵盖目标确定、任务分层、人机协同与多维评价的作业设计框架，并在工程类课程中开展实证研究。研究分三阶段展开，分别对应工科学生知识习得、能力形成与素养内化的进阶训练。经过程性数据与结果性数据的统计分析发现，参与实证研究的学生对复杂工程概念的理解与方案设计能力均有显著提升，学生交互模式的聚类研究结果呈现分层特征，皮尔逊相关性研究结果显示交互模式与学习效果呈显著正相关。未来教学可进一步探索按交互模式层级实施差异化分层的引导策略，以促进各类工程学生实现深度协同成长。

Abstract: In view of the current situation that traditional engineering assignment cannot support the cultivation of complex engineering problem solving capabilities, this study explores a way for generative artificial intelligence to improve assignment design. Taking synergy theory as the theoretical foundation to clarify the design ideas of human-machine primary-secondary relationships and nonlinear interaction, and using scaffolding theory to guide the progressive design of assignment task gradients and human-machine interactive feedback, this study constructs an assignment design framework covering goal determination, task stratification, human-machine collaboration, and multidimensional evaluation, and conducts empirical research in engineering course instruction. The study unfolds in three stages, corresponding respectively to the progressive training of engineering students in knowledge acquisition, ability formation, and literacy internalization. Through statistical analysis of process data and outcome data, it is found that students participating in the empirical research show significant improvement in their understanding of complex engineering concepts and their scheme design ability. The clustering research results of students’ interaction patterns present stratified characteristics, and Pearson correlation results indicate a significant positive correlation between interaction patterns and learning outcomes. Future instruction may further explore differentiated, stratified guidance strategies implemented according to interaction pattern levels, so as to promote deep collaborative growth among engineering students of all types.