[英語教學資源中心] 沒有界限的界限　AI與EMI教學輔助工具工作坊

國立政治大學英語教學資源中心邀請到美利堅大學 （American University）全球與沉浸式研究辦公室（Office of Global and Immersive Studies Faculty）Professorial Lecturer Krisztina Domjan，於5月27日早上蒞臨本校法學院演講廳進行「AI與EMI教學輔助工具工作坊」（AI-n’t a Given: Mastering Key AI Literacy Skills and Pedagogical Techniques）演講。除了實體參與的觀眾，這次演講也透過網路同步直播。本場活動由英語教學資源中心主任鍾曉芳開場，她指出這次講座的英文標題裡的「AIn’t」有雙重意義——除了AI以外，在英文裡的「ain’t」一字，其實是「is not」或「are not」的縮寫——也就是說，人工智慧並非理所當然；我們必須謹慎並充分認識地使用它，此乃AI素養（AI literacy）所需。

這次演講的最大特色為互動性與即時性——如同Krisztina Domjan在課堂裡的教學，會預先將演講的簡報上傳至互動平台，並在投影片裡設計互動式問題。出席者除了可以透過電腦、平板或手機觀看簡報，並且點閱附在投影片的連結以外，還可以同時現場留言作出回應或表達意見，而她也把觀眾的發表同時顯示在演講廳的屏幕上，與觀眾進行互動。

Krisztina Domjan表示自己並非AI擁護者，也不是資料科學家（data scientist）。其實，她在AI聊天系統於2022年底推出時，才開始試驗這些系統並研究其背後的教育價值。她期望透過使用AI，可以節省執行各項任務的時間，就像她的學生也利用AI作為完成作業的捷徑。因此她深信教育工作者，應當學習如何在自身的教學中有效地引用AI，而學生也要學會根據個人需要適當地使用AI聊天系統，使其成為自己的免費朋友或助手，讓學習過程變得更加容易。為了強調AI的重要性，她引用了字母控股（Alphabet）及Google執行長Sundar Pichai的名言：「比起火、電力，還有我們過去所做的任何事，AI都來得更有深度」（AI is more profound than fire, electricity, or anything that we have done in the past）。

這次演講分為三個部分：一、AI素養以及教學中AI工具與聊天系統的概述；二、AI科技所牽涉到的評估與道德兩難；三、在教學中結合AI的策略。根據Krisztina Domjan所述，這次演講的基本概念是AI素養，簡單來說，就是我們需要具備理解、使用和評價AI的能力，包括認知其相關原理，如應用、可能的影響、限制與道德考量。AI工具在我們周圍唾手可得，它們可分為三大類別：

1. 自適應AI（Adaptive AI）：例如語言學習應用程式Duolingo，這類AI通過適應使用者的需求並與使用者進行互動來學習。

2. 助理式AI（Assistive AI）：這類AI可以幫助使用者在他們的文檔或Google Chrome中工作。

3. 生成式AI（Generative AI）：這類AI聊天系統能夠為使用者生成文字、影像甚至編碼。

講者進一步介紹了一些對於語言學習有用的助理式AI工具。例如，Announcify和Speechify能夠將文字轉換成語音，使你能夠聆聽文章；另一方面，在Google或MS Word的文檔中，Voice Typing則可以將語音轉換成文字，節省打字的時間。此外，Postlight Reader能夠讓你在線上閱讀文章時，專注於主要內容，而Rewordify及NoteGPT則能夠簡化複雜的文章或影片，使所有讀者都易於理解。這些工具都能夠幫助學習者在處理大量材料與資訊時，有效地節省時間與精力。

在眾所周知的生成式AI工具中，Krisztina Domjan認為ChatGPT非常有價值，但如果想要生成標題，Gemini則是更好的選擇。此外她強調了建立和評估指令的重要性：指令越清晰，所產出的結果就越好。她鼓勵觀眾在使用AI聊天系統時提供清晰的指令，並附上相關背景信息。她指出了「審問生成結果」的重要性：特別是在無法獲得預期結果時，必須對指令進行修改和反覆嘗試。為了使AI的產出更加個性化，產出應該根據個人特徵進行調整。總之她期望使用者能夠對AI的產出進行批判性思考。

Krisztina Domjan在演講中討論了另一個教育工作者普遍關心且無法忽視的重要概念：學生作業中可以包含多少由AI生成的內容。她建議教授、課程主任和院長應共同討論以下問題：如果學生的作業中包含了AI協助生成的內容，他們應該如何表述？在課程和作業中，教師應包括什麼引述指引？除了參考文獻頁面（references page）以外，是否需要包含材料和方法的章節（Materials and Methods section）？教師在教室中對相關界限所做的意圖和期望必須公開透明。

以Krisztina Domjan的課堂為例，她採用了五個等級的尺度，讓學生了解在作業中使用AI的程度：未使用AI、AI協助構思與結構、AI協助修改、在人類檢閱下由AI完成的特別任務、完全使用AI而不讓人類察覺。此外，她與學生討論什麼是適合的AI工具，以及在什麼情況下使用AI可能導致作弊。她鼓勵學生使用她在演講前部分介紹的工具，而不是在作業中加入ChatGPT、Gemini、Copilot和Claude.ai，或其他生成式AI工具所生成的內容，因為這些內容都會被檢測出來。學生也需要理解誠信的概念，當他們在寫作中使用AI工具時，需要聲明並附上hyperdoc。總體來說，對課程設計和教學指導進行修改，以及透明度都是不可或缺的。這樣可以使學生清楚地知道在使用AI時，什麼是可以接受的，什麼是不可以接受的，從而避免產生疑問。

出席者不僅在整場演講中表現出熱情和專注，還有一部分觀眾在演講結束後留下來向講者尋求更多指導。英語教學資源中心對Krisztina Domjan提供的詳盡而實用的貢獻表示感謝。她的貢獻讓我們了解到如何在課堂中使用更多的AI聊天系統進行教學和評量。

＃更多資訊：

請上官網　www.eminccu.com

FB　https://www.facebook.com/nccu.emi

IG　https://www.instagram.com/nccu.emi 

The EMI Resource Center at National Chengchi University invited Dr. Krisztina Domjan, Professorial Lecturer from the Office of Global and Immersive Studies Faculty at The American University, to give an inspirational talk entitled “AI-n’t a Given: Mastering Key AI Literacy Skills and Pedagogical Techniques.” The talk took place in the Lecture Hall of the College of Law, NCCU, on the morning of Monday, May 27th, 2024. Besides the attendance of the audience in the Lecture Hall, this event was also simultaneously broadcast online. Professor Siaw-Fong Chung, the Center’s director, opened the event by explaining the split inside the title of this talk – AI, beyond doubt, is the main theme of the talk, while “ain’t” is the short form of “is not” or “are not”—that is, AI is not a given; we must use it with caution and acknowledgement. That is what AI literacy needs.

This talk featured interaction and immediacy. Dr. Domjan, as she often does in her classroom teaching, uploaded her presentation slides online. In several of the slides, she posed questions throughout the talk. Attendees could access the slides, which included various links, on the spot using a code provided by the speaker. They could type their responses or opinions, which were simultaneously displayed on the Lecture Hall’s screen, through their laptops or cellphones.

Dr. Domjan started the talk by stating that she was neither an AI evangelist nor a data scientist. She only began experimenting with AI and studying its educational value after the release of AI chatbots in late 2022, hoping to save time on tasks using AI, similar to how her students use AI as shortcuts for homework. Dr. Domjan believes educators should learn how to incorporate AI technologies into their pedagogy seamlessly, while learners should learn how to customize AI chatbots as unpaid friends or assistants. These were the dual aims of her talk. She emphasized the significance of AI with a quote from Sundar Pichai, CEO of Alphabet and Google: “AI is more profound than fire, electricity, or anything that we have done in the past.”

The talk was divided into three parts: first, AI literacy and an overview of AI tools in education and AI chatbots; second, AI technologies in assessment and the ethical dilemmas involved; and finally, pedagogical strategies for integrating AI. According to Dr. Domjan, AI literacy involves the ability to understand, use, and evaluate AI, including its applications, implications, limitations, and ethical considerations. AI tools can be categorized into adaptive AI, such as Duolingo, which learns from user interactions; assistive AI, which extends usage to documents or browsers like Google Chrome; and generative AI, which creates text, images, or code.

Dr. Domjan introduced several assistive AI tools for language learning, such as Announcify and Speechify (text-to-speech tools), and Voice Typing in Google Docs or MS Word (speech-to-text tools). Postlight Reader helps focus on online articles, while Rewordify and NoteGPT simplify complex texts or videos. These tools save learners time and effort in managing learning materials. Among generative AI tools, she highlighted ChatGPT and recommended Gemini for title creation. Dr. Domjan stressed the importance of constructing and evaluating prompts: clear instructions and context lead to better outputs. She advised adapting and iterating instructions to achieve desired results and encouraged users to build upon AI outputs with personal input, maintaining a critical approach.

Another key topic was the acceptance of AI-generated content in student assignments. Dr. Domjan suggested faculty should collaboratively address questions about AI assistance and citation in assignments. She proposed transparency in the classroom regarding AI use, with a 5-level scale to inform students about the allowed extent of AI use in assignments. This scale ranges from no AI use to full AI use with human oversight. She also recommended early adoption of specific AI tools rather than generative AI for assignments and emphasized teaching integrity by requiring disclaimers when AI tools are used.

The audience's enthusiasm was evident throughout the talk and afterward, with many staying to seek further advice from Dr. Domjan. The EMI Resource Center expressed heartfelt gratitude to Dr. Krisztina Domjan for her informative and practical contributions, which expanded understanding of how AI chatbots can be effectively utilized in language teaching and assessment.

For more information, please check our website：

www.eminccu.com

FB：https://www.facebook.com/nccu.emi

IG：https://www.instagram.com/nccu.emi