← → Arrow keys or click buttons

AI in Mathematics
Education

Possibilities and Challenges

How People Learn Mathematics in the Age of Artificial Intelligence

Pedagogical Insights · Cognitive Strategies · Inclusive Learning

Panel: Teacher Education in the Digital and AI Era

Pedagogy · Competency · Inclusive Learning

Context: Nepal | Higher Education

🧮

References: Chomsky (2023) · Bloom (1956) · Dehaene (2011) · Piaget (1952) · Vygotsky (1978) · Nepal NEB 2078

What is AI? — Three Perspectives

1

AI as Simple Machine

⚙️ Processes data and follows rules. Calculates, sorts, searches. No understanding — fast computation only.

A calculator that solves 12+15=27 without knowing what addition means.

2

AI as Necessary Tool

🔧 A useful instrument — like a pencil or textbook. Teachers and students use it to enhance learning, not replace thinking.

Chomsky (2023): AI is a tool shaped by human language and thought — but not a thinker itself.

3

AI as Sufficient Answer?

⚠️ The dangerous view: that AI alone can replace teachers and education. This misunderstands both learning and teaching.

"Teaching is a science and art — so a teacher cannot be coded."

How Do People Learn?

→ 12 + 15 = 27

①

① Count Forward

15 + · · · → 27

Working memory counts on from the larger number. Procedural, high cognitive load.

Piaget (1952)

②

② Decompose & Add

10+10=20 · 2+5=7 → 27

Place-value split. Conceptual understanding of base-10 structure.

Vygotsky (1978)

③

③ Make a Ten

12+8=20 · 15−8=7 → 27

Compensatory strategy. Flexible number sense. Early algebraic thinking.

Number sense research

④

④ Spatial Mapping

12 →+10→ 22 →+5→ 27

Mental number line. Humans naturally map numbers to space.

Dehaene (2011)

⑤

⑤ Formal Operation

Column: ones then tens

Standard algorithm. Procedural fluency at formal operational stage.

Piaget formal stage

⑥

⑥ Machine / Recall

Direct retrieval from memory

Automaticity frees cognitive load for higher reasoning. Goal of fluency practice.

Cognitive load theory

✦ Possibilities: AI in Mathematics Education

🎯

Personalised Learning Paths

AI adapts to each student's pace and cognitive level. Khan Academy's AI tutor gives individualised feedback — serving both advanced and struggling learners simultaneously.

Holmes et al. (2019)

📐

Visualising Abstract Concepts

GeoGebra and Desmos (AI-enhanced) let students see number lines, geometric transformations, and algebraic relationships interactively — anchoring abstract maths in spatial intuition.

Dehaene (2011)

⚡

Immediate Feedback & Error Analysis

AI identifies the exact cognitive strategy a student used and where it broke down. Not just "wrong" — but WHY and HOW to correct it. Transforms assessment into teaching.

Black & Wiliam (1998)

🌍

Inclusive & Accessible Mathematics

Text-to-speech, multilingual AI, and assistive tools bring mathematics to students with disabilities, remote learners in Humla and Dolpa, and non-Nepali speakers alike.

UNESCO (2023)

🏛️

Teacher as Learning Designer

AI handles drill and recall, freeing teachers to focus on higher-order thinking, discussion, and creative problem-solving. Teacher becomes architect of learning experience.

Chomsky (2023)

⚠ Challenges: The Dangers We Must Face

🧠

AI Cannot Think Like a Human

Chomsky (2023): ChatGPT is "a kind of super-autocomplete" — generates statistically likely text, not genuine reasoning. AI has no feelings, no curiosity, no moral compass. It cannot wonder WHY mathematics matters.

CRITICAL

🎮

Illusion of Learning

Students who use AI to get answers do not build cognitive schema. Gamification without reflection creates the feeling of learning without its substance. The brain must struggle to grow.

PEDAGOGY

📡

Digital Divide in Nepal

In Humla, Jumla, and Dolpa: no electricity, no internet, no devices. AI-powered maths in Kathmandu deepens inequality with rural Nepal. Infrastructure must precede pedagogy.

EQUITY

📚

Curriculum Misalignment

Nepal's B.Ed. and M.Ed. programmes (NEB 2078) include no AI literacy or computational thinking. Teachers graduate without competency to critically evaluate AI tools in mathematics classrooms.

POLICY

⚖️

Ethical Risk: Who Does AI Serve?

AI trained on Western datasets embeds cultural, linguistic, and racial bias. Nepali, Maithili, Tharu, and Tamang learners are underrepresented. Algorithms may systematically disadvantage marginalised students.

ETHICS

Human vs AI — Never Competitors, Always Partners

👩‍🏫

HUMAN TEACHER

✦ Wonders, questions, and feels genuine curiosity

✦ Understands context — a child's fear, a family's struggle

✦ Makes ethical and moral judgements

✦ Teaches with empathy and relationship

✦ Adapts to the unseen emotional reality

✦ Inspires a student to love mathematics

✦ Asks: "Why does any of this matter?"

+

partner

🤖

AI TOOL

✦ Processes millions of problems instantly

✦ Gives personalised, patient feedback 24/7

✦ Adapts difficulty to individual learner

✦ Identifies patterns in student errors at scale

✦ Makes mathematics accessible across languages

✦ Handles routine drill — freeing teacher for depth

✦ Generates: "Here is the next step..."

✦ Lesson: Teach students to "Work WITH AI" — not to be replaced by it, and not to fear it.

🎮 Teaching Critical Thinking Through Gamification

REMEMBER

Flashcards, quizzes — points for correct recall

UNDERSTAND

Drag & drop, sorting — explain to unlock next level

APPLY

Mission simulations — use maths to solve a real crisis

ANALYSE

Detective puzzles — find the flaw, spot the pattern

EVALUATE

Debate arenas — score others' arguments with evidence

CREATE

Build, design, make something new — peer review

🔑 Ensuring Real Learning

Require explanation before advancing — not just the answer

Switch student role: What? How? Why?

Answer additively — scaffold understanding upward

Portfolio over leaderboard score

Peer teaching to "certify" mastery

⚠ Nepal Context

B.Ed./M.Ed. currently: Remember ↔ Performance only

Must shift to Create/Design level thinking

Offline gamification works for Humla, Dolpa

AI cannot ask "Why does any of it matter?"

✓ This Is Learning

Game doing → question → answer additive → THIS IS LEARNING

Engagement ≠ Learning. Make thinking the price of admission.

Can AI Think? — Chomsky's Insight Applied to Mathematics

"

ChatGPT and similar programs are... a kind of super-autocomplete — systems designed to produce the statistically most likely output for a given input. Such a system cannot distinguish truth from falsehood or right from wrong; and no commitment to honesty or morality can be derived from it.

— Noam Chomsky, New York Times, March 8, 2023

🤖 What AI does in mathematics

→ Generates statistically likely correct answers

→ Pattern-matches from training data

→ Produces step-by-step solutions on demand

→ Has no understanding of why 12+15=27 matters

→ Cannot experience mathematical beauty or curiosity

→ Ask 5 AIs the same prompt → similar reactions

🧠 What human thinking adds

✦ Genuine understanding — not just output

✦ Curiosity: "I wonder what happens if..."

✦ Moral judgement: "Should we use this formula?"

✦ Creativity: new mathematical ideas and proofs

✦ Meaning: "Why does any of it matter?"

✦ Ask 5 people the same prompt → different reactions

What Must We Teach? — Skills for the AI Era Workplace

💻

AI & Digital Literacy

Use AI tools critically. Understand bias, limitations, prompting, data literacy, basic computational thinking.

Every 2030+ workplace uses AI. Students who can't use it will be left behind.

🧠

Critical Thinking

Evaluate AI outputs. Ask: Is this true? Is this fair? Is this the best solution? Mathematical reasoning beyond calculation.

AI can generate answers — humans must judge which are good, true, and ethical.

❤️

Human Skills (EQ)

Communicate, collaborate, empathise. AI cannot replace human connection, care, or moral leadership in the classroom.

Empathy, care, and human connection cannot be automated.

⚖️

Ethics & Responsibility

Every student will face AI decisions. A moral compass is a core graduate competency — not an optional extra.

AI without ethics is a danger. Students need values, not just skills.

🔄

Lifelong Learning

AI literacy + human wisdom = ability to keep learning. The skill of learning HOW to learn is the most future-proof skill.

Jobs of 2040 don't exist yet. Learn to learn — always.

🇳🇵

Nepal AI Literacy Context

AI skills + ethical values + human skills. Communicate + collaborate + moral compass. B.Ed./M.Ed. must embed these.

AI cannot replace teachers — but teachers must understand AI to guide students wisely.

Education Redefined — From "Copy" to "Create"

THEN — Old Education

Educare: to mould

✗ Copy the teacher

✗ Memorise the answer

✗ Reproduce knowledge

✗ One right answer

✗ Teacher as authority

✗ Student as empty vessel

✗ Learn once, use forever

NOW — AI Era Education

Educere: to draw out

✦ Question, explore, discover

✦ Find the right question

✦ Create new knowledge

✦ Multiple possible solutions

✦ Teacher as guide and co-learner

✦ Student as active thinker and leader

✦ Learn continuously — adapt always

"Old education: 'Here is the answer. Remember it.'  |  New education: 'Here is a problem the world has never faced. What do you think?'"

🇳🇵 Recommendations for Nepal: Mathematics in the AI Era

01

📋

Reform B.Ed. & M.Ed. Curriculum

Embed AI literacy, computational thinking, and digital pedagogy. Move beyond remember ↔ performance to create/design thinking.

02

🎮

Bloom-Based Gamification

Design mathematics learning games that require explanation, not just correct answers. Switch student role: What? How? Why?

03

🏔️

Offline-First for Remote Districts

Humla, Jumla, Dolpa need infrastructure before apps. Physical gamification, debate tournaments, and community problem-solving work without electricity.

04

🗣️

Nepali Language AI Development

Invest in Nepali, Maithili, and Tharu NLP. AI tools that serve only English speakers deepen exclusion. Mathematics must be linguistically inclusive.

05

🌱

Teacher as Learning Architect

Teaching is a science and art — a teacher cannot be coded. AI handles drill; teachers inspire. Reposition teacher as guide, not knowledge-deliverer.

✦

💡

Core Insight

AI cannot ask "Why does any of it matter?" — but teachers and students of Nepal can, and must.

Closing Thought

"AI can generate. AI can calculate. AI can imitate.
But AI cannot wonder, care, or ask —
'Why does any of it matter?'
That is the teacher's irreplaceable gift to mathematics education."

References

• Chomsky, N. (2023). The False Promise of ChatGPT. New York Times, Mar 8.
• Dehaene, S. (2011). The Number Sense. Oxford University Press.
• Piaget, J. (1952). The Child's Conception of Number. Routledge.
• Vygotsky, L. S. (1978). Mind in Society. Harvard University Press.
• Bloom, B. S. (1956). Taxonomy of Educational Objectives. Longmans.
• Holmes, W. et al. (2019). Artificial Intelligence in Education. CCR.
• Black, P. & Wiliam, D. (1998). Inside the Black Box. Phi Delta Kappan.
• UNESCO (2023). AI Competency Framework for Teachers. Paris.
• Nepal NEB (2078 B.S.). Teacher Education Curriculum. CDC.

1 / 12