Primary school students’ perceptions of artificial intelligence – for good or bad

University department (environmental and life sciences / science education)

Karlstad, Sweden

Researcher facilitating data collection and analysis

• In-school (K–12)

Case study: pre-test, focus group interviews, post-lesson evaluations

Short instructional sequence with reflective components

Primary cohort aged 11–12 (classroom case)

Students explore various AI tools (incl. post-ChatGPT awareness)

• Single case generalization limits
• Children’s rights framing requires careful facilitation
• Tool landscape evolves quickly after data collection
• Self-reported use may under/over-state

11–12 (Swedish primary)

Rising ChatGPT-era public awareness

Not emphasized

• Map cognitive and affective perceptions of AI
• Document if/how students already use AI tools
• Connect findings to child rights and regulation debates

• Ground future AI literacy curricula in student voice
• Inform policymakers listening to children

• Not national representative survey
• Not longitudinal multi-year tracking
• Not technical competency test

General AI tools referenced by students (e.g., assistants, ChatGPT-era apps)

• Tutor
• Automation tool

Swedish schooling context

• Students explore tools then reflect in groups
• Ethical and societal concerns surface spontaneously

• School rules for minor use of GenAI
• Privacy education for children
• Balanced framing of labor and dystopia fears
• Adult moderation of tool exploration

• Administer pre-test on AI conceptions
• Run focus groups on good/bad perceptions and usage
• Collect post-lesson evaluations
• Fuse Mitcham dimensions with AI literacy pillars in analysis

• Students voice preferences for slowing AI via regulation; adults translate into policy learning

• Philosophical framework scaffolds abstract “what is AI?” discussions

• Tension between helpful study support and societal worries
• Need regulations that children can understand and trust
• Cognitive categories (machine/concept/human-like) vary

• Integrated philosophical framework with behavioural use component for richer coding

• Students engage seriously with ethics and futures

• Mixed affect: optimism plus privacy and job concerns

Positions children’s voices as foundation for AI literacy and policy in education.

• UNCRC-aligned listening and assent
• Avoid sensationalizing while acknowledging real risks
• Secure handling of child-generated writing about AI
• Equitable access so exploration is not privilege-based

• Post assessment
• Activity documentation
• Practitioner observation

• Larger multilingual studies of child AI perceptions
• Link perceptions to measured digital citizenship competencies

• Primary science education
• AI literacy
• Children’s rights

• Completed

11–12

Sweden

Perception + early use

Case study / reflection

Low–Medium

Medium