Artificial Intelligence education for young children: Why, what, and how in curriculum design and implementation

University faculty of early childhood education

Hong Kong SAR, China

Single-author synthesis with literature grounding

• In-school (K–12)
• Informal learning

Exploratory literature review informing curriculum principles and exemplar unit design

Conceptual paper (not a timed intervention)

Ages 3–8 target cohort in design recommendations

AI toys, embodied apps, code.org-style activities referenced in exemplar framing

• Scarce empirical ECE AI curriculum evidence at time of writing
• Risk of overstating mechanistic understanding for youngest learners
• Digital divide affects access to quality AI learning experiences
• Teacher capacity for interdisciplinary STEM+AI integration

Young children approximately 3–8 years

High informal exposure to assistants and media; uneven guided instruction

Not required; emphasis on exploration over formal coding

• Justify early AI education as digital citizenship and equity
• Define a feasible subset of ML-related ideas for early years
• Specify embodied, culturally responsive pedagogies for inquiry with AI tools

• Link AI education to broader STEM integration
• Position ethical and limitation-aware use from the start

• Not a randomized trial of learning gains
• Not a single-tool efficacy study
• Not exhaustive of all global ECE policies

Age-appropriate AI interfaces, robots, and microworlds (exemplar-oriented)

• Tutor
• Co-creator

Context-dependent; paper anchored in international ECE discourse

• Playful interaction with intelligent agents
• Teacher-mediated framing of data, training, and failure modes
• Culturally situated inquiry tasks

• Mitigate misleading AI outputs with adult scaffolding
• Protect child data and vendor transparency
• Avoid deficit framing of families with less device access
• Balance wonder with honest limits of models

• Ground rationale in digital equity and intelligent society trends
• Select concepts children can experience without heavy formalism
• Sequence embodied, relational activities (exemplar “AI for Kids”)
• Reflect on cultural responsiveness and teacher facilitation

• Educators curate tasks and ethics; AI systems illustrate patterns
• Children explore; adults interpret errors and fairness

• Learning-by-making with tangible and screen-based bridges
• Relational pedagogy emphasizing dialogue and joint attention

• Tension between rapid consumer AI spread and uneven school readiness
• Need for curricula that do not assume prior programming
• Equity gaps in who gets high-quality AI learning experiences

• Explicit Why–What–How curriculum structure for ECE audiences
• Ties AI literacy to existing digital literacy and STEM policy conversations

• Synthesis cites prior work showing promise of robots and playful ML for young children

• Core early idea: models learn from data to predict/act with limitations

Offers a pedagogical model and exemplar pathway rather than new empirical outcome data from a single cohort.

• Child-safety and accuracy when using conversational or recommender AI
• Equitable access to devices and high-quality facilitation
• Transparent consent for any data-generating classroom tools
• Age-appropriate honesty about automation and bias

• Activity documentation
• Practitioner observation

• Empirical evaluations of “AI for Kids” style progressions across cultures
• Longitudinal equity studies linking early AI encounters to later critical evaluation skills

• Early childhood technology education
• AI literacy and developmental appropriateness
• Culturally responsive STEM

• Completed

3–8

ECE classrooms / informal

Introductory ML patterns + ethics

Embodied inquiry, learning-by-making

Medium

Medium