From "Learning-Application Disconnect" to "Learning for Application": An Exploration of the Reform Path for General Artificial Intelligence Courses

Abstract

To address the "learning-application disconnect" in university general AI courses, this paper analyzes pain points such as outdated content, monotonous teaching methods, and wide disparities in students' foundations. It proposes a three-dimensional objective of "knowledge— ability—literacy" and three principles: "low code + high understanding," "scenario-based + project-based," and "disciplinary intersection." The reform path is explored through content reconstruction, blended teaching, and diversified assessment. Practice shows that course satisfaction rose from 80% to 97%, students' confidence in AI application increased from 54% to 89%, and the project completion rate reached 100%, providing a replicable paradigm for the inclusive development of general AI education.