Teach AI from Year 1, not just Year 13, urges digital education expert

AI in Schools: Start earlier, build capacity, keep it broad

The Government plans to roll out new secondary school subjects from 2028, including a specialised Year 13 course on generative AI. The intent is clear: put more weight behind science, technology, engineering, and maths.

Experts support the move but warn the current approach starts too late, narrows the focus, and risks running into a teacher capacity wall. Here's what needs to happen now to make the policy work.

Start before Year 13

Canterbury University associate professor of digital education Kathryn MacCallum is blunt: waiting until Year 13 is too late. Many students already use AI tools well before senior secondary school.

AI literacy should begin in primary years and build through Year 13. That means a clear progression that grows with students' maturity and subjects.

• Years 1-6: Simple concepts (what AI is and isn't), safe use, recognising misinformation, basic privacy.
• Years 7-10: Data basics, how recommendations work, limits and bias, prompt quality, citing AI assistance.
• Years 11-13: Generative AI use in projects, fundamentals of machine learning, cybersecurity practices, ethics, and when not to use AI.

Teach the system, not just the tools

Focusing only on generative tools misses the larger picture. Students should learn how digital systems and machine learning work, how AI can mislead, and where AI has no place in an academic or civic process.

MacCallum's point is practical: AI literacy should connect to digital citizenship. That means explicit guidance on appropriate use, critical thinking, consent, and accountability across subjects.

Teacher capacity is the choke point

Former secondary teacher and Teachers Institute founder Dr Nina Hood flags the core risk: there aren't enough specialist teachers to deliver the new content. Building capacity must start now.

• Fund national professional development with release time and micro-credentials.
• Partner with industry for guest experts, externships, and co-designed projects.
• Stand up regional hubs to share lesson plans, assessments, and exemplars.
• Provide baseline guidance on classroom AI use, academic integrity, and privacy.

Guard against unintended consequences in other subjects

Mount Aspiring College principal Nicola Jacobson warns that shifting outdoor education to a "vocational" track and removing its University Entrance status risks narrowing students' choices and devaluing important learning. Clear binary lists (academic vs vocational) can create stigma and limit pathways.

• Keep multiple routes to University Entrance that reflect diverse strengths.
• Review credit and assessment settings so practical subjects aren't sidelined.
• Communicate with parents and students to prevent misperceptions about value.

A practical rollout plan to hit 2028

• Now-2026: Define AI literacy standards across Years 1-13; publish teaching guidance and integrity rules; seed pilot schools.
• 2026-2027: Scale teacher upskilling; co-develop projects with industry; build assessment exemplars for NCEA, including cross-curricular work.
• 2028 launch: Introduce the Year 13 generative AI subject and embed AI literacy across subjects; fund ongoing support and evaluation.

Policy decisions to make this stick

• Make AI literacy explicit from Year 1 and integrate it into digital citizenship.
• Broaden beyond tools: systems thinking, data, ethics, cybersecurity, and clear "no-go" zones for AI use.
• Invest in teacher capacity with time, training, and partnerships.
• Align assessment and University Entrance rules to avoid devaluing practical subjects.

Resources

• Ministry of Education: Curriculum and Assessment Changes
• Complete AI Training: Curated AI courses by job role

Bottom line: Starting AI at Year 13 won't be enough. Begin early, teach the full system, fund the workforce, and protect breadth across the curriculum.