Scholarlink

<p>This study investigates the role of ChatGPT, a generative artificial intelligence tool, in enhancing learning within computer science and engineering education. Conducted during the Fall 2023 semester, the work utilized ChatGPT based on the GPT-3.5 model to assess its effectiveness across three educational domains: programming, analytical problem-solving, and hardware-related system design. A structured methodology was implemented wherein students from multiple academic levels attempted tasks independently and then used ChatGPT for support. Their interactions and reflections were systematically recorded and analyzed to evaluate the model’s utility and limitations in an educational setting. Results indicated that ChatGPT was particularly effective in programming tasks, where it could generate accurate code and useful explanations. However, its performance declined in more complex areas such as artificial intelligence and computer architecture, where deeper analytical reasoning and domain-specific expertise were required. The study further highlights the value of training students in prompt engineering, which significantly improved their ability to formulate targeted queries, increased metacognitive awareness, and helped calibrate their trust in the tool’s responses. Recommendations include future integration of graphical diagram support to aid comprehension in visually complex subjects, such as system design and architecture. Key challenges identified include limited transparency in training data, the lack of visual support, and the need to assess whether prompt engineering skills and improved AI usage persist and transfer across courses. While newer reasoning-oriented models such as GPT-4, Claude 3, and DeepSeek may offer enhanced capabilities, these were not available during the Fall 2023 study period and are suggested for future evaluation. The study also calls for future research using these newer, more advanced large language models to reassess findings and ensure continued relevance as these tools evolve. Overall, this research provides valuable insights into the educational application of generative AI and informs future strategies for integrating AI tools into engineering pedagogy.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Effectiveness of Generative AI Tools in Computer Science and Engineering Education

  • 期刊论文

  • 发表日期：

    2025年10月23日

查看全文
  • Maria Waqas,
  • Shehzad Hasan,
  • Anita Ali

摘要

This study investigates the role of ChatGPT, a generative artificial intelligence tool, in enhancing learning within computer science and engineering education. Conducted during the Fall 2023 semester, the work utilized ChatGPT based on the GPT-3.5 model to assess its effectiveness across three educational domains: programming, analytical problem-solving, and hardware-related system design. A structured methodology was implemented wherein students from multiple academic levels attempted tasks independently and then used ChatGPT for support. Their interactions and reflections were systematically recorded and analyzed to evaluate the model’s utility and limitations in an educational setting. Results indicated that ChatGPT was particularly effective in programming tasks, where it could generate accurate code and useful explanations. However, its performance declined in more complex areas such as artificial intelligence and computer architecture, where deeper analytical reasoning and domain-specific expertise were required. The study further highlights the value of training students in prompt engineering, which significantly improved their ability to formulate targeted queries, increased metacognitive awareness, and helped calibrate their trust in the tool’s responses. Recommendations include future integration of graphical diagram support to aid comprehension in visually complex subjects, such as system design and architecture. Key challenges identified include limited transparency in training data, the lack of visual support, and the need to assess whether prompt engineering skills and improved AI usage persist and transfer across courses. While newer reasoning-oriented models such as GPT-4, Claude 3, and DeepSeek may offer enhanced capabilities, these were not available during the Fall 2023 study period and are suggested for future evaluation. The study also calls for future research using these newer, more advanced large language models to reassess findings and ensure continued relevance as these tools evolve. Overall, this research provides valuable insights into the educational application of generative AI and informs future strategies for integrating AI tools into engineering pedagogy.