In the era of proliferating artificial intelligence (AI) technology, generative AI is reshaping educational landscapes, prompting a critical examination of its influence on students' learning processes and their self-efficacy amid concerns over growing technological dependence. This study investigates the nuanced relationship between generative AI use and university students' self-efficacy and technological dependence, illuminating the underlying paradoxes and implications for inclusive education practices. Through a survey of 348 university students, with 200 valid responses analyzed, we uncover the direct and indirect impacts of generative AI usage frequency on AI dependence. Our findings reveal a paradoxical effect: enhanced AI usage significantly amplifies students' confidence and efficiency in learning, yet simultaneously intensifies their dependence on AI. This dual impact both supports and complicates the incorporation of AI technologies into educational settings, underscoring the need for a balanced approach to leveraging AI in teaching and learning. Our study underscores the critical importance of a nuanced understanding of AI's role in education. It highlights the necessity of crafting an educational landscape where technology augments learning processes without compromising independent learning capabilities. By navigating the complex interplay between technological advancement and educational inclusivity, our findings guide the development of AI-assisted learning environments that are not only effective but also equitable and accessible.

中文翻译：

---

自我效能感和技术依赖的悖论：揭示生成式人工智能对大学生任务完成的影响


在人工智能 （AI） 技术激增的时代，生成式 AI 正在重塑教育格局，在对技术日益增长的依赖的担忧中，促使人们批判性地审视其对学生学习过程及其自我效能感的影响。本研究调查了生成式 AI 使用与大学生自我效能感和技术依赖之间的微妙关系，阐明了潜在的悖论和对包容性教育实践的影响。通过对 348 名大学生的调查，分析了 200 个有效回复，我们揭示了生成式 AI 使用频率对 AI 依赖性的直接和间接影响。我们的研究结果揭示了一个矛盾的效果：增强的 AI 使用显着增强了学生的学习信心和效率，但同时加剧了他们对 AI 的依赖。这种双重影响既支持将 AI 技术纳入教育环境，又使其复杂化，凸显了在教学中利用 AI 的平衡方法的必要性。我们的研究强调了细致入微地理解人工智能在教育中的作用的极端重要性。它强调了精心打造教育环境的必要性，在这种环境中，技术可以增强学习过程，而不会影响独立学习能力。通过驾驭技术进步和教育包容性之间的复杂相互作用，我们的研究结果指导开发不仅有效而且公平且可访问的 AI 辅助学习环境。