Keynote Address

Professor Rose Luckin

UCL Institute of Education

Founder, EDUCATE Ventures Research

About the Speaker

We are honoured to welcome Professor Rose Luckin as our keynote speaker. Prof Luckin is the Professor of Learner Centred Design at University College London (UCL) and a globally recognised leader in the field of Artificial Intelligence in Education. She is the author of AI for School Teachers and the founder of EDUCATE Ventures Research, which bridges research, policy, and practice to support effective use of AI and educational technology. With her pioneering work on AI literacy and preparing learners for the future, Prof Luckin brings invaluable insights for educators navigating the opportunities and challenges of AI in teaching and learning.

Title of sharing

Human-Centred AI and the Future of Learning: Building Meta-Intelligence in Education

Abstract

As artificial intelligence reshapes every aspect of society, education stands at a critical crossroads. Will we allow AI to diminish human intelligence, or will we harness it to create “super intelligence” the powerful combination of human and artificial capabilities working in synergy?

This keynote addresses the urgent need to move beyond simply integrating AI tools to fundamentally reimagining how we develop human intelligence in our students. Through the “Three Lenses of AI in Education” framework, I’ll explore: AI as a subject we must teach, AI as a tool that can enhance learning, and AI as infrastructure that can capture sophisticated thinking.

Drawing on recent neuroscience research showing concerning declines in human reasoning when over-relying on AI, I’ll argue that our priority must be developing “meta-intelligence” the sophisticated human capabilities of metacognition, critical thinking, and self-regulated learning that enable students to collaborate effectively with AI rather than depend on it passively.

Educators will gain practical strategies for:

Building AI literacy that goes beyond tool use to critical evaluation
Redesigning assessment and pedagogy for an AI-augmented world
Ensuring ethical AI use
Developing the uniquely human skills students need to thrive
Implementing learning analytics that capture and nurture complex thinking

Rather than fearing AI or banning it, we must embrace our evolving role as educators: not as content deliverers but as architects of learning experiences that build human capabilities AI cannot replicate. This is education transformation that truly serves human flourishing in the age of artificial intelligence.

Presentation Deck

2. [Keynote] Singapore Nov 2025 Human-Centred AI Building Meta-Intelligence in Education.pptx

Keynote Address Video

Keynote Address - Prof Rose Luckin.mp4

Question and Answer

Q1. What specific safeguards or practices would you recommend to ensure 17-18 year old students develop robust metacognitive skills rather than becoming overly dependent on AI scaffolding?

Key Recommendations:

Progressive Fading: Systematically reduce AI support across the term; include "AI-free" versions of previously assisted tasks
Mandatory Reflection: Students document AI interactions in learning journals - "What did AI suggest? Why did I accept/reject it? Could I do this alone?"
Strategy-First Approach: Teach metacognitive strategies (planning, monitoring, reflection) BEFORE introducing AI tools
Mixed Assessment Design: Alternate between AI-assisted and traditional assessments so students compare their performance and strategy use
Critical Evaluation Tasks: Students identify when AI suggestions are wrong or inappropriate, building awareness of AI limitations

Core Principle:
AI should amplify existing metacognitive skills. For university-bound students, the goal is strategic AI users, not AI-dependent learners.

Q2. Do we have casual evidence on how AI affects less motivated students, beyond task completion, to retention/transfer and self regulation? Which assessments or scaffolds reduce passive reliance?

Honest Answer:
We have surprisingly little causal evidence specifically examining different motivation profiles for generative AI.

What We Know:

Less motivated students likely complete more tasks WITH AI
BUT we lack evidence on retention, transfer, and self-regulation development
Risk: AI becomes a permanent crutch rather than temporary scaffold

Scaffolds That May Reduce Passive Reliance:

Process-Focused Assessment: Credit for documenting problem-solving process, not just answers
Graduated Prompting: AI provides hints only after student attempts independent strategies first
Pre-AI Metacognitive Prompts: Students must answer "What do I know? What's my plan?" before AI access
Delayed Feedback with Self-Assessment: Students evaluate their own work before seeing AI feedback
Collaborative AI Use: Students work in pairs, explaining thinking to peers before consulting AI

Bottom Line:
Until we have robust causal evidence, adopt a precautionary approach—design AI scaffolds that require active engagement and monitor quality of thinking, not just task completion.

Critical Research Gap:
We urgently need longitudinal studies examining how different AI scaffolds affect motivation, retention, transfer, and self-regulation across diverse student populations.

Remember: "Learn Fast, Act More Slowly"

Page updated

Report abuse