Human-Centered AI in Schools

The OSPI Human-Centered AI in Schools framework (see link here:https://ospi.k12.wa.us/student-success/resources-subject-area/human-centered-artificial-intelligence-schools#:~:text=Educational%20policymakers%20must%20focus%20on,and%20strengthening%20learning%20management%20systems.) emphasizes the ethical, inclusive, and responsible use of AI tools in education, focusing on supporting both teachers and students.

What is “Human-Centered AI in Schools”?

The idea behind Human-Centered AI (HCAI) in education is to ensure that as AI tools (like chatbots, AI writing tools, adaptive learning systems) enter classrooms, the focus stays on students, teachers, and ethical learning processes — not on letting AI dictate how education happens. It means designing and using AI that aligns with human values, equity, fairness, and educational goals.

Key Principles from Human-Centered AI in Schools (based on OSPI and broader research):

1. AI should augment, not replace, human teachers.

AI tools should support teachers in teaching more effectively, not take over their role.

Example: AI that helps give personalized feedback but leaves final decisions to the teacher.

2. Students remain at the center — AI should empower, not control.

AI should enhance students’ creativity, critical thinking, and agency, not make them passive users.

Example: AI that helps students generate ideas for writing, but still requires them to think, reflect, and revise.

3. Equity and fairness.

AI should be used to support diverse learners (e.g., neurodivergent, dyslexic students), not widen gaps in education.

Example: AI that adapts reading materials to student levels, including accessible formats.

4. Transparency and ethics.

Teachers and students should understand how AI tools work, what data they collect, and how to use them safely.

Example: Explaining to students what AI can and cannot do, and why teachers may choose to use or not use certain AI tools.

5. Critical AI literacy for teachers and students.

Students and teachers should learn how to critically assess AI-generated content, question biases, and reflect on ethical use.

Example: Teaching students to analyze AI-generated Mandarin sentences for accuracy and naturalness — “Does this make sense in real-life conversation?”

What Teacher education and Mandarin teaching Can Do:

1. Training Mandarin teachers to use AI ethically and effectively.

How can AI tools support vocabulary, pronunciation, and character learning without replacing authentic teacher-student interaction?

How to critically assess AI-generated materials (e.g., ChatGPT translations that might be wrong or awkward).

2. Supporting neurodivergent learners using AI.

AI tools (e.g., text-to-speech, adaptive quizzes) can support dyslexic or ADHD learners in Mandarin, but teachers need to know how to choose and adapt these tools carefully.

3. Developing teachers’ AI literacy — not just how to use the tools, but how to think critically about AI in education.

Helping teachers evaluate AI outputs, question biases, and stay updated with fast-evolving technology.

4. Interdisciplinary and critical thinking approach — fits beautifully with your idea of encouraging teachers to be creative, research-informed, and reflective.

Draft PGCE Mandarin Module — “AI in the Mandarin Classroom: Human-Centered, Inclusive, and Effective”

Example AI-Enhanced Classroom Activities for Mandarin Teaching

Activity 1: AI Chatbot Role-Play for Real-Life Situations

Purpose: Practice conversational Mandarin in simulated real-life contexts (e.g., ordering food, asking directions).

Tool: ChatGPT or HelloTalk AI chatbot.

Task:

• Students interact with AI chatbot to complete a scenario (e.g., ordering at a restaurant).

• Reflect on chatbot responses — are they natural, culturally appropriate?

• In pairs, students revise chatbot dialogue to make it more authentic.

Teacher role: Discuss how to adapt chatbot language for real conversations; focus on correcting AI’s unnatural output.

Activity 2: Speech Recognition for Pronunciation and Tone Practice

Purpose: Improve tone accuracy and fluency.

Tool: Speechling or HelloChinese.

Task:

• Students record themselves reading a dialogue.

• AI tool gives feedback on tones and pronunciation.

• Teacher helps analyze AI feedback: “What did AI miss? What do you think is more accurate?”

Differentiation: Extra visual tone markers for dyslexic students, or allow repetition at their pace.

Activity 3: Collaborative AI-Aided Writing Task

Purpose: Practice character writing, sentence-building, and grammar.

Tool: ChatGPT (writing prompts), Pleco (dictionary support).

Task:

• Students use AI to generate simple sentences based on a theme (e.g., “My family”).

• In groups, students review and correct AI sentences.

• Rewrite sentences using classroom vocabulary, ensuring cultural appropriateness.

Assessment option: Students record themselves reading their final texts as speaking practice.

Activity 4: AI for Differentiated Vocabulary Support (Neurodivergent Learners)

Purpose: Help dyslexic or ADHD learners with vocabulary retention.

Tool: Quizlet AI, Skritter, color-coded pinyin generators.

Task:

• AI generates vocabulary flashcards with pictures and sound.

• Students group vocabulary by radicals or themes (e.g., family words, food).

• Create personalized vocabulary sets using AI support — students choose what helps them (visual, audio, color cues).

Activity 5: Ethical AI Reflection Task

Purpose: Develop critical AI literacy in both teachers and students.

Task:

• Compare an AI-generated dialogue to a real native-speaker conversation — what are the differences?

• Discuss: “Should we trust AI’s answers? How do we know if AI is wrong?”

• Students write a reflection: “If I use AI to help me learn Mandarin, what do I need to watch out for?”

Add-On: Teacher Resource Bank

Invite trainee teachers to create and share AI-enhanced lessons and reflections in a shared space (Google Drive, Padlet, Notion).

• Categories: Speaking, writing, inclusive teaching, assessment, cultural learning.

• Grow into a living toolkit for future cohorts.

Note: Parts of this blog were written with the assistance of ChatGPT, an AI language model developed by OpenAI. The final content was edited and reviewed by me to reflect my personal voice and perspective.