As artificial intelligence reshapes workplaces faster than curricula can keep up, schools should shift from teaching fixed toolsets to building learners’ ability to adapt, argue educators working at the frontier of vocational secondary education.
From specific skills to durable capabilities
The founder and director of Armenia’s ACT College — a state-licensed vocational high school focused on technology and the creative industries — says five years is now “an eternity in technology”. The school admits learners from about age 15 and offers distinct learning tracks that aim to cultivate practical, resilient capabilities rather than mastery of a single, soon-outdated tool.
“The goal is not simply for them to understand what artificial intelligence is, but to understand how it works.”
At ACT College, learners select from tracks in computer science, engineering and manufacturing, or digital art. In the computer science stream, the sequence moves from mathematics into Python, algorithms and machine learning. But the emphasis is on using AI to build and iterate, learning from failure and understanding why a solution does not work — not on packaging a fixed set of skills and declaring a learner ready for the workplace.
Work-based projects and the changing role of qualifications
The school reports that some learners are already engaged in real-world projects while still at school. That practice reframes qualification systems: rather than certifying that a learner has memorised a set curriculum, education systems should recognise the capacity to learn new tools, to troubleshoot AI-driven systems and to participate in collaborative problem-solving.
That approach raises practical questions South African policymakers face as well: how to structure teacher training so instructors can guide projects with rapidly changing technology; how to align assessment and certification with adaptive capabilities; and how to ensure access to the necessary equipment and digital infrastructure in poorer schools.
- Teach for adaptability: prioritise meta-skills such as critical thinking, experimentation and learning-to-learn.
- Embed hands-on projects: create sustained opportunities for learners to build and iterate with AI tools.
- Reconsider assessment: move beyond static tests to evaluations that show learners can apply and adapt skills in new contexts.
Why this matters for South African classrooms
The argument is not to discard knowledge but to change priorities. In a world where a tool learned today may be obsolete in a few years, success will depend on learners’ ability to understand underlying principles, to test and correct their work, and to translate competence across different platforms and job roles.
For South Africa, which faces persistent inequality in school resources, the shift to capability-led learning will require targeted investment in teacher development and in access to project-grade equipment and connectivity. Without that, the gap between well-resourced schools and those in under-served communities may widen as employers demand skills shaped by exposure to real-world, technology-rich tasks.
| ACT College track | Core focus |
|---|---|
| Computer science | Mathematics, Python, algorithms, machine learning; building AI projects |
| Engineering & manufacturing | Practical systems and production-focused skills (reported at the school) |
| Digital art | Creative industry techniques blended with digital tools |
The table reflects the reported structure of learning tracks at ACT College and highlights an organising principle: combine domain knowledge with project-based experience so learners develop transferable capabilities.
Longer term planning in this context means designing curricula and qualifications that accept uncertainty. Rather than trying to predict the exact tools learners will use in five years, systems should prepare learners to identify new tools, evaluate them and apply them within existing knowledge frameworks.
Education leaders and policy-makers must therefore ask hard questions about assessment design, teacher development and resource allocation. For parents and communities the immediate takeaway is clear: the most valuable outcomes may not be a polished certificate showing mastery of a single software package, but evidence that a learner can tackle unfamiliar problems, fix errors and meaningfully contribute to technology-enabled work.
Adapting to an AI-driven labour market is less about chasing the latest tool and more about ensuring learners leave school able to learn again.
