> ## Content Index
> Fetch the complete content index at: https://katecarruthers.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# AI in schools: Australia should choose learning before automation
- URL: https://katecarruthers.com/ai-in-schools-australia-should-choose-learning-before-automation/
- Published: 2026-10-18T20:24:38.000Z
- Updated: 2026-10-18T20:24:38.000Z
- Description: AI is already entering Australian classrooms. The question is whether it will support learning, or allow students to outsource their thinking.
- Author: Kate Carruthers
- Tags: ai, education, students, children, Australia, digital sovereignty, technology policy, schools, generative ai

AI is already entering Australian classrooms, whether schools are ready or not. The question is not whether students will use it. The question is whether schools will let it substitute for learning, or teach students how to understand, challenge and use it without outsourcing their thinking.

[Norway](https://www.csmonitor.com/Editorials/the-monitors-view/2026/0424/Scandinavia-s-U-turn-on-book-reading?ref=katecarruthers.com) and China offer two very different models. One starts with protecting learning and child development. The other starts with building national AI capability. Australia should learn from both, but copy neither. Our approach needs to start with children, teachers, public education and the infrastructure we are willing to depend on.

## The Nordic instinct: protect learning

Norway has taken a deliberately cautious approach. From the 2026 school year, generative AI is generally prohibited for students aged 6 to 13; students aged 14 to 16 may use it cautiously under teacher supervision; and older secondary students are expected to learn appropriate use in preparation for work and further study, according to [Reuters](https://www.reuters.com/technology/norway-imposes-near-ban-ai-elementary-school-2026-06-19/?ref=katecarruthers.com).

The logic is straightforward: early schooling is where children develop foundational literacy, numeracy, concentration, problem-solving and the capacity to express their own thinking. If an AI system completes the hard part of a task too early, it can remove precisely the productive struggle through which learning occurs.

This is not an argument for preserving the classroom in amber. It is an acknowledgement that a seven-year-old does not need an AI writing assistant. They need to learn to read, write, calculate, draw, ask questions and persist when something is difficult.

Norway’s decision also sits alongside a renewed emphasis on physical books in classrooms. That matters. Schools should be places where students have access to technology, certainly, but not places where screens and automated systems displace every tactile, social and cognitive part of learning. [GovTech’s overview of international approaches](https://www.govtech.com/education/k-12/how-poland-norway-and-uae-are-approaching-ai-in-schools?ref=katecarruthers.com) describes the wider policy shift.

## The Chinese approach: build national capability

China has approached AI education as a [national capability-building](https://english.www.gov.cn/news/202604/15/content%5FWS69df29e6c6d00ca5f9a0a6b1.html?ref=katecarruthers.com) project. Its Ministry of Education issued guidance in late 2024, followed by detailed 2025 material designed to establish a tiered AI education system across primary and secondary schooling. The stated aim is to develop innovative talent and national capacity for future technological challenges, as outlined in the [Australian Government’s overview of China’s approach](https://www.education.gov.au/international-education-engagement/resources/chinas-approach-ai-education-primary-and-secondary-schools?ref=katecarruthers.com) and [Chinese government material](http://english.scio.gov.cn/pressroom/2025-05/13/content%5F117871666.html?ref=katecarruthers.com).

The approach is structured by developmental stage. Students progress from basic awareness and understanding of AI, through practical applications, towards design, innovation, ethics and social responsibility. Beijing’s implementation plan includes a minimum of eight AI education hours each year and allows schools to teach it as a dedicated subject or integrate it across existing subjects.

There are sensible elements here. AI literacy cannot simply be left to chance, or to whichever students have the most digitally confident parents. A national framework can give schools a common baseline, support teacher training and ensure that questions of ethics, privacy and social responsibility are taught alongside technical capability.

But China’s model is also deeply strategic. It links school education to national technological competitiveness and a future AI workforce. Australia should understand that context before importing the model wholesale. Education is not simply a talent pipeline for the technology sector. It is where a democracy develops people who can read critically, reason independently and challenge systems that claim to be objective.

## Handwriting is not obsolete

One casualty of the rush towards AI-enabled learning could be handwriting. It should not be.

Handwriting is not simply an old-fashioned way of producing text. For developing writers, it requires the coordination of spelling knowledge, letter forms and fine motor movement while also drawing on attention and working memory. As handwriting becomes more automatic, children can devote more mental effort to planning, generating ideas and revising what they write.

A [2022 systematic review and meta-analysis](https://pmc.ncbi.nlm.nih.gov/articles/PMC9204270/?ref=katecarruthers.com) of 31 studies involving 2,030 students found that handwriting interventions produced significantly better writing fluency than no handwriting instruction or instruction unrelated to handwriting. The review reported an overall effect size of 0.64.

That is not an argument against keyboards, speech-to-text or assistive technology. Those tools are essential for many students and useful for everyone. It is an argument against confusing access to polished output with the development of the underlying capability to think, compose and communicate.

A child who cannot yet form a sentence independently should not be given an AI system that can form paragraphs on demand. The technology may produce words, but it cannot develop the child’s fine motor control, writing fluency, vocabulary, judgement or confidence in their own voice.

## Australia should copy neither model

- [**The Nordic approach**](https://www.reuters.com/technology/norway-imposes-near-ban-ai-elementary-school-2026-06-19/?ref=katecarruthers.com) **starts with child development.** Norway’s restrictions on generative AI for younger students recognise a basic fact: children need to develop literacy, numeracy, concentration, problem-solving and their own capacity to express an idea before we hand them a machine that can produce plausible answers on demand.
- [**The Chinese approach**](https://www.education.gov.au/international-education-engagement/resources/chinas-approach-ai-education-primary-and-secondary-schools?ref=katecarruthers.com) **starts with national capability.** China is building AI education across the school years, from basic awareness through to application, innovation, ethics and social responsibility. It treats AI literacy as a strategic capability, not an optional extra.
- **Australia should copy neither model.** We should not leave students unprepared for a world shaped by AI. But nor should we organise schooling around the workforce needs, commercial interests and technological assumptions of a handful of AI companies.
- **Teachers need to remain in charge.** The question is not whether students have access to a chatbot. The question is whether teachers have the time, training, curriculum resources and authority to decide when a tool helps learning, and when it simply does the learning for the student.
- **Australia already has a framework, but frameworks are not implementation.** The [Australian Framework for Generative AI in Schools](https://www.education.gov.au/schooling/resources/australian-framework-generative-artificial-intelligence-ai-schools?ref=katecarruthers.com) provides a useful foundation for responsible and ethical use. Education Ministers endorsed its 2024 [review](https://www.education.gov.au/schooling/announcements/2024-review-australian-framework-generative-artificial-intelligence-schools?ref=katecarruthers.com) in June 2025\. The harder work is translating those principles into classroom practice, procurement standards and meaningful support for teachers.
- **The Australian position should be straightforward.** Teach children *about* AI before expecting them to use it. Protect reading, handwriting, discussion, practical work and independent drafting in the early years. Then introduce supervised AI use gradually, with transparency, verification and human judgement at the centre.

## A sensible Australian model

Australia should adopt an age-based, capability-based approach rather than a blanket embrace or ban.

### Early primary: no generative AI for student work

For the early years, schools should prioritise reading, handwriting, number sense, oral language, play, practical activities and teacher-led learning. Generative AI should not produce student work, and it should not become a substitute for teachers, books or classroom relationships.

Children can still learn simple concepts: that computers follow rules, that data is collected, that automated systems can make mistakes, and that people need to check what technology does. That is AI literacy without handing a chatbot to a six-year-old.

### Upper primary: teach critical literacy

From around Years 5 and 6, students can begin to examine AI outputs under close teacher guidance. They should learn that systems can invent information, reproduce bias, expose private data and present confident answers that are wrong.

This is also the right age to teach the difference between using an AI system to explore an idea and submitting AI-generated work as if it were their own. The issue is not only cheating. It is authorship, judgement and intellectual independence.

### Secondary school: supervised use and verification

Secondary students need experience using AI because they will encounter it in university, workplaces and public life. But the task design needs to change.

Students should be required to:

- Show their reasoning, drafts and source selection.
- Verify claims made by AI tools against reliable sources.
- Explain which parts of a task were assisted by AI.
- Identify errors, bias and missing perspectives in generated output.
- Use AI for critique, simulation, coding support or alternative explanations, not simply for first-draft substitution.
- Consider privacy, copyright, security, energy use and the commercial incentives behind the tools.

This is more demanding than banning AI or allowing it without constraint. It is also much closer to the capability Australian students will need.

## Schools need control, not just tools

The central issue is not whether a school has access to a chatbot. It is whether teachers have the time, training, curriculum resources and authority to decide when AI use helps learning and when it harms it.

Australia should fund:

- Nationally available, curriculum-aligned AI literacy resources.
- Professional learning for teachers, school leaders and support staff.
- Clear guidance on assessment, privacy, child safety and procurement.
- Public-interest evaluation of classroom AI tools before large-scale adoption.
- Accessible assistive technologies for students who benefit from them.
- Procurement rules that prevent schools from casually sending student data to offshore commercial platforms.

The Australian Government’s analysis of [China’s approach to AI education in schools](https://www.education.gov.au/international-education-engagement/resources/chinas-approach-ai-education-primary-and-secondary-schools?ref=katecarruthers.com) notes the importance of teacher training, age-appropriate curricula, project-based learning and ethics alongside technical skills. Those are useful principles, even if Australia should pursue them through our own democratic, educational and privacy frameworks.

School AI is also a data governance issue. Every AI platform used in a classroom raises questions about where student and teacher data is stored, whether prompts are retained, how data is used to train models, whether the service can be audited, and what happens when a vendor changes its terms.

Schools should not become an easy source of children’s data for offshore technology platforms. The questions are practical: where does the data go, who can access it, can the system be independently audited, which jurisdiction applies, and can a school or department walk away if a vendor changes its terms? If those questions cannot be answered, the tool is not ready for the classroom.

## The test Australia should apply

Before introducing AI into a classroom, schools should ask one simple question:

**Does this use of AI deepen learning that a student could not reasonably achieve otherwise, or does it remove the learning we are meant to be developing?**

If it removes the learning, then do not use it.

If it supports accessibility, provides useful feedback, exposes students to real-world tools or helps them interrogate how automated systems work, it may have value. But it should be introduced deliberately, transparently and in ways that preserve human judgement.

Australia does not need to choose between a nostalgic rejection of technology and an education system organised around AI platforms. We can protect childhood, teach handwriting and foundational knowledge, build serious AI literacy, and prepare students to participate in a world where automated systems shape work, government and daily life. That is the balance worth getting right.