By Willem Kitshoff
The question is no longer whether artificial intelligence belongs in South African classrooms. It is already there, formally through approved school platforms, and informally through learner usage and teacher experimentation.
The real challenge is moving beyond the hype to provide schools with practical clarity, consistency, and implementation frameworks that let them use these technologies effectively and responsibly.
As Riaan van der Bergh, Deputy CEO of the Federation of Governing Bodies of South African Schools (FEDSAS), recently observed, the priority is giving teachers and learners the exact frameworks they need to navigate these tools safely. That is not a technological hurdle alone. It is fundamentally a challenge of governance, operational readiness, and system design.
Schools are among the most complex institutions in society. They simultaneously juggle teaching, communication, administration, finance, legal compliance, and pastoral care. In that high-pressure environment, any new technology must reduce complexity, or it risks compounding it.
Two decades of working with schools have taught us one simple truth: efficiently run schools are often the best-performing ones. When communication flows smoothly and administrative burdens lift, educators can do what they are there to do: teach. That lesson is critical as AI enters the fold.
While public conversation fixates on AI’s capabilities, the more important question is integration. AI can support personalised learning, streamline administration, and free educators from routine tasks. But without strict oversight, it introduces serious risks: misinformation, fragmented systems, and data privacy breaches.
The policy gap in our schools is widening. Well-intentioned teachers are uploading learner work, report data, and behavioural notes into public generative AI platforms to save time, inadvertently exposing personally identifiable information to public algorithms and raising immediate flags under the Protection of Personal Information Act (POPIA). The answer is not to ban innovation. It is to govern it.
Schools need clear guidance on what data can and cannot be shared with AI systems. Governing bodies need policies that remove legal uncertainty. And schools need technology partners who understand the unique legal responsibilities that come with managing learner information.
This is the philosophy behind our AI Assist development at d6. We have taken a deliberate, layered approach. The first layer, embedded in d6 School Communicator, uses AI to answer parent and community queries, drawing exclusively on verified, school-approved content and operating within a POPIA-aligned environment where the school retains full oversight.
The next layer goes deeper. AI Assist is now embedded across the d6 school management platform, giving staff, from administrative and curriculum teams to finance and SGB leadership, direct conversational access to their own institutional data.
Rather than navigating complex reports or waiting for extracted summaries, staff can ask questions in plain language and receive immediate, contextualised answers from within their secure environment. The system meets people where they are, in the language they use, while keeping all data governed, role-restricted, and within the school’s control. That is the difference between AI as a shortcut and AI as a properly integrated institutional tool.
We are also excited about what comes next. The natural progression beyond AI that answers questions is AI that acts. We are actively building agentic AI capabilities within the d6 platform, which will allow schools to move from insight to action, automating routine tasks, triggering workflows, and enabling staff to delegate defined administrative processes to AI agents operating within governed, school-approved boundaries. The goal is not to replace human judgement, but to free up the people who hold it.
There is also a significant equity risk at play. Well-resourced schools can afford to develop independent AI policies and train staff. Millions of learners in under-resourced schools cannot. Without a coordinated national approach, AI adoption will widen the digital divide rather than close it.
South Africa faces a clear choice: let AI adoption happen ad hoc, driven by fragmented experimentation, or build the governance frameworks, training programmes, and trusted systems required to elevate the entire sector. The latter demands urgent collaboration between government, governing bodies, educators, and technology providers.
Technology alone does not improve education. Effective, safe implementation does.
We do not need more debate about whether AI belongs in education. Reality has answered that. What we need now is the structural discipline to ensure it strengthens our schools rather than complicates them. The challenge is no longer one of possibility. It is one of readiness.
Willem Kitshoff is Chief Executive Officer of d6, a leading South African school management and communication technology provider.
INSIDE EDUCATION
