In the context of technological innovation driving the development of educational technology, collaborative problem-solving (CPS) and innovation skills have become essential core competencies for high-quality talent. The importance of collaborative problem-solving in personal development has also garnered increasing attention from society. However, current educational efforts to develop students’ collaborative learning and problem-solving skills mainly rely on teacher-led classroom instruction. From the perspective of the integration of digital and intelligent technologies, this article innovatively shifts the focus to the cultivation of human-AI collaborative learning abilities. It specifically examines how students engage in human-AI collaboration using ChatGPT technology within different knowledge contexts across various learning stages. This study tracks students’ mastery of current knowledge points and their collaborative learning processes with ChatGPT, using paired t-tests, independent t-tests, ANOVA, and qualitative feedback analysis to evaluate the impact on their collaborative problem-solving skills. The findings demonstrate that human-AI interaction plays a significant and valuable role in improving students’ collaborative learning abilities, providing substantial support for educators in future efforts to cultivate and enhance students’ collaborative problem-solving skills in teaching.
This study systematically reviews the feasibility of dialogue learning based on generative AI. Research on mixed reality and generative AI in enhancing teaching effectiveness is increasingly in-depth both domestically and internationally. The development and application of mixed reality and generative AI are accelerating, evolving classroom structures and methods. In the context of rapid social, economic, and technological development, students require more personalized and innovative education, which dialogue learning interventions based on generative AI can effectively address. Through a systematic literature review method, 128 relevant studies from the past decade were systematically analyzed to explore the feasibility of dialogue learning in the context of generative AI. The study finds that the application of generative AI in dialogue learning not only enhances learning outcomes but also excels in personalized learning and timely feedback. However, the diversity and operability of technological implementation and application scenarios require further exploration. This research provides references and insights for future studies by related scholars.
This study addresses the challenge of extreme weather on property insurance, proposing a model that balances insurer profitability and affordability for owners. By integrating strategic underwriting, normalized data, and advanced forecasting, the model aims to enhance system resilience and close the Insurance Protection Gap, particularly in high-risk areas like the U.S. and Europe, which are increasingly affected by natural disasters. To achieve these goals, we employ state-of-the-art computer technologies, including 3D scene analysis for enhanced visualization and assessment of property risks, as well as machine learning algorithms for time series forecasting. These techniques, along with least squares optimization and robustness checks, allow for more accurate risk evaluation and support the development of sustainable underwriting practices.