應用ChatGPT於行動裝置助理系統開發—以AR雙語學習遊戲為例

外語學習需要長時間的練習與對話才能熟練，當學習者在外語不熟練和不熟
悉的狀況下會畏懼用外語與人溝通。然而，近年來人工智慧的發展迅速，能透過
人工智慧去識別對話並給予回覆。學習者能透過與人工智慧進行外語交流，來訓
練外語交流的能力。本研究將建立一款擴增實境雙語學習遊戲，透過與遊戲的助
理系統互動來進行實驗。
本研究藉由助理系統去引導學習者去進行外語對話，應用 OpenAI推出的
ChatGPT 生成對話內容，並透過擴增實境在現實環境中模擬對話場景，讓助理系
統去引導學習者對話。並透過易用性測驗去測試助理系統能否有效地輔助學習者
進行外語交流。本研究藉由與學校課程活動搭配，讓學習者遊玩擴增實境雙語學
習遊戲，並填寫易用性問卷和使用者感受，根據結果分析成效與提供未來建議。
在本次的測驗中雖然結果不如預期，但能看到結合擴增實境的助理系統對使
用者而言有一定的吸引力，對於助理系統操作上的瑕疵也獲得修改的意見，也將
針對未來助理系統獨立於遊戲外的互動方式提出意見。

Learning a foreign language requires prolonged practice and conversation to
become proficient. When learners are unfamiliar or unskilled in a foreign language, they
may fear communicating with others in that language. However, with the rapid
development of artificial intelligence in recent years, it is now possible to identify
conversations and provide responses through AI. Learners can engage in foreign
language communication with AI to train their language skills. This study aims to
develop an augmented reality bilingual learning game and conduct experiments by
interacting with the game's assistant system.
The assistant system in this study will guide learners in foreign language
conversations by using OpenAI's ChatGPT to generate dialogue content. Through
augmented reality, the system will simulate dialogue scenarios in a real-world
environment, guiding learners through conversations. Usability tests will be conducted
to assess the effectiveness of the assistant system in aiding learners' foreign language
communication. By integrating this game with school curriculum activities, learners will
play the augmented reality bilingual learning game and complete usability questionnaires
and user feedback forms. The results will be analyzed to evaluate the effectiveness and
provide future recommendations.
Although the test results did not meet expectations, the integration of the augmented
reality assistant system showed a certain level of appeal to users. The feedback also
highlighted operational flaws in the assistant system, providing valuable suggestions for
future improvements. Additionally, recommendations for future interaction methods for
the assistant system outside the game context will be proposed.