● Emerging international research into the use of generative AI in education has found that AI can support learning when used with clear teaching principles, directly supporting the distinction between producing a convincing answer and demonstrating real understanding.
● According to the World Economic Forum’s Future of Jobs Report 2025, 69% of employers consider analytical thinking an essential core skill, with creative thinking considered essential by 57%, and technological literacy identified by 51%. This reinforces the need for schools to develop students who can use new technology effectively while continuing to think critically, creatively and independently.
● The University of Cape Town’s wide approach towards AI places greater emphasis on the process of learning, including oral assessment, observed group work and students critically reflecting on their use of AI. This indicates that parts of the local education sector are beginning to move beyond detection towards reconsidering how students demonstrate genuine understanding.
September 2026: Seven out of 10 employers regard analytical thinking as an essential core skill, according to the World Economic Forum’s Future of Jobs Report 2025. 57% of employers consider creative thinking essential, while 51% identify technological literacy as a core skill. Young people will need to use technology confidently, but also with judgement, creativity and independence.
The OECD’s Digital Education Outlook 2026 makes the educational case equally clearly: generative AI can support learning when guided by sound teaching principles. But when students use it simply to outsource a task, it may improve the work they produce without delivering genuine learning gains. A convincing answer, in other words, is not always evidence of understanding.
AI is already part of how many students search, write, revise, code and create. The question is no longer whether it belongs in education. It does. The question is how schools ensure it supports learning rather than replaces the thinking students need to do themselves.
At Koa Academy, the response is neither unrestricted AI use nor blanket prohibition. It is the AI Compass: an internally developed framework and set of teaching guidelines that gives students, parents and teachers a shared language for responsible AI use.
The Compass is integrated into coursework, tasks and assessments, making the expected use of AI explicit for each activity. It outlines four modes: Independent, Explorer, Collaborator and Producer.
Independent tasks require students to complete the work without AI. Explorer tasks allow AI to help students investigate ideas or clarify difficult concepts, while the final work remains their own. Collaborator tasks use AI for feedback and refinement, such as reviewing writing, debugging code or testing an argument. Producer tasks invite students to use AI to create a final product, requiring them to prompt, direct, select and edit the output.
“The AI Compass is not a blanket permission to use AI, and it is not a blanket ban. It gives students clear guidance for each task: should they work independently, use AI to explore a concept, collaborate with it to improve work they have started, or use it intentionally to create a final product?” says Mark Anderson, Co-founder of Koa Academy.
This specificity is central to the approach. A vague instruction that AI may be used does not explain what meaningful and ethical use looks like. Equally, a blanket ban does not equip students to make informed decisions about a technology they will encounter in higher education, work and daily life.
A task might, for example, specify: “Collaborator: Before you submit the final task, use Gemini to suggest where you can improve your paragraph construction and grammatical structure.” The instruction identifies the permitted role of AI and the student’s continuing responsibility for the work.
The Compass has been introduced through teacher training, parent engagement and student sessions. Responsible AI use is a capability that must be practised repeatedly across subjects and contexts.
Schools and universities are right to be concerned about plagiarism and academic integrity. Clear rules are necessary. But rules and detection tools alone will not prepare young people for an AI-enabled world.
South Africa is already seeing a shift in this conversation. The University of Cape Town discontinued Turnitin’s AI-detection score in October 2025, citing concerns about its reliability and fairness. Its approach instead places greater emphasis on the learning process, including oral assessment, observed group work and students’ critical reflection on AI use.
When AI can produce fluent text, solve formulaic problems and generate plausible answers in seconds, the completed product becomes less reliable as evidence of learning. The deeper question is not only, “Did the student use AI?” It is also: “What did the student need to think through, explain, question or apply to complete this task?”
“A polished final answer is not always proof of learning. We need to pay closer attention to how students approach a task, the decisions they make, the questions they ask, and whether they can explain and apply what they have learnt,” says Anderson.
At Koa Academy, teachers assess learning as it occurs, not only through a final submission. This places importance on the process behind the outcome: the reasoning, decisions, information and tools a student used and whether they can defend and apply their learning.
It also deliberately develops critical thinking, communication, ethical judgement, curiosity, collaboration, problem-solving and leadership. These are not additions to academic learning. They are developed through how students learn, participate, apply knowledge and work with others.
“AI makes these skills even more important. When producing an answer becomes easier, the real differentiator is whether a young person can evaluate that answer, improve it, apply it and use it responsibly,” explains Anderson.
The AI Compass recognises that AI can be a powerful tool for exploration, feedback, practice and creation, while protecting the thinking and demonstrated understanding that make learning meaningful. Koa Academy’s goal is not to prepare students for a world without AI, but for one in which they can use it confidently, critically and ethically, while retaining ownership of their thinking, learning and ideas.
For more information about Koa Academy, visit https://koaacademy.com.
