1. Background & Motivation 

I chose this course because I wanted to improve my project management competencies and logical thinking skills. Prior to this course, one of the recurring academic difficulties I faced was dealing with computing problems. As an English major, I initially possessed the preconception that AI tools would be hard for me to handle. I was genuinely curious and concerned about whether I could catch up with the curriculum, given that my background knowledge of computers and AI applications was strictly limited.

2. Digital Humanities & Project Description 

Our project bridges cultural literacy with interactive technology, focusing on a comparative analysis of cultural rituals surrounding remembrance: the traditional Taiwanese Ghost Festival (中元節) versus the Western Day of the Dead (西方亡靈節). Through this cross-cultural exploration, we aim to introduce elementary school students to international perspectives while grounding them in the unique values of Chinese and Taiwanese culture.

The core implementation of our project involves the Kebbi Robot (凱比機器人), where we wrote specialized programs to make the robot interact dynamically with young learners. To catch the students' eyesight and keep them highly engaged, Kebbi interacts with the children through a multi-modal system: voice-based Q&A (語音問答), touchscreen selections (觸碰螢幕), and manual typing of answers (手動輸入答案). Through these interactive sessions, students can advance their listening skills and language knowledge while watching videos and stories.

3. AI Collaboration & Prompt Modification

In my project, AI mainly acted as a vital supporter to help me deal with tasks much quicker and more effectively. I used Gemini, NotebookLM, and Ed-cafe to dig out creative ideas and design interesting interactive segments, while relying on the Roboflow examples provided during class to ensure the technical accuracy of my workflow.

However, we encountered significant technical friction when testing the robot's interaction sequence. During oral question-and-answer phases, the robot frequently exhibited a listening disability, failing to accurately capture the children's spoken responses. The initial AI output regarding the interaction flow was awful, requiring us to test and revise the design many times.

To solve this problem, my method of interacting with AI underwent a major transformation. At the beginning, I merely provided superficial prompts, simply asking the AI for general feedback regarding my project ideas (e.g., "What do you think about my current interactive lesson design for Kebbi?"). To break through the roadblock, I changed my approach and shifted toward collaborative brainstorming and deeper conceptual scaffolding. I began asking AI much more detailed and specific questions to address our technical constraints (e.g., "The Kebbi robot is failing to recognize oral inputs from elementary students due to ambient noise. Help me redesign the interaction script so that if oral recognition fails twice, the screen automatically prompts a multi-choice touch menu or manual input field. Provide the logic flow for this."). In addition to this prompt refinement, I also actively sought out my teammates' technical support and advice to successfully resolve these programmatic glitches.

4. Human Agency & Quality Control 

The human component of our project was strongly demonstrated through my own creativity, pedagogical direction, and original ideas. Although AI assisted in generating the initial story patterns, character voices, and dialogue templates, I did not simply accept the AI outputs passively. Instead, I collected my own instructional thoughts first, and then relied on AI strictly to refine the final results and make them better. Furthermore, I systematically verified the AI's responses by searching authoritative websites to ensure that all cultural and language information was factually accurate and perfectly fit my educational expectations.

5. Academic Innovation & Practical Value

Compared with traditional language learning methodologies, our project stands out because the interactive robotic sessions transform passive studying into an active experience. The most innovative part of our project is the integration of multi-tiered interaction mechanisms (voice, touch, and input) combined with captivating questions designed to hold the students' attention.

The most practical value of this project is learning how to design engaging story patterns and simple English structures to help elementary students advance their learning. In a real classroom, this project could be utilized to foster a deeper understanding of cultural differences and cross-cultural empathy. For self-directed learning, primary students—the target audience for whom this simple English script was tailor-made—can use this project to boost their listening comprehension, while igniting a genuine passion for global cultural events.

6. Reflection, Learning Shifts, and Limitations (心得與學習轉變)

At the beginning of the semester, I thought AI was a difficult and intimidating tool to implement in daily life, but now I recognize that it vastly improves our time management and increases our ability to consider issues from multiple perspectives. My biggest "Aha moment" occurred when we successfully navigated these technical hurdles; I realized that completing a digital humanities project was not as insurmountable as I had previously thought. I discovered that I learn much better when AI summarizes dense materials, which ultimately frees up my cognitive bandwidth for critical thinking.

I have truly learned that working with AI does not mean outsourcing our brains or thinking less. Instead, it means exploring new possibilities to co-create with AI to reduce mechanical workload and achieve superior results. One major limitation of AI is that it is incredibly difficult to satisfy all of our nuanced academic needs in only one single prompt. Therefore, students must be highly cautious not to treat AI as infallible or discard our own critical thinking. In the future, I can use AI as an autonomous agent to handle even more complex projects, but I will always make sure to verify the accuracy of the information and continuously improve my personal capabilities, ensuring I remain the master of my own craft rather than a prisoner of creativity.

Guiding Reading

Presentation to know further about metacognition (MA) and chatGPT