Training one million people – but for what?

South Africa is promising massive digital and AI skilling drives for public servants and citizens, but unless training is tied to real job and systems change, we risk producing certificates instead of capability.

Article link: https://open.substack.com/pub/johanosteyn/p/training-one-million-people-but-for?r=73gqa&utm_campaign=post&utm_medium=web&showWelcomeOnShare=true

Audio summary: https://youtu.be/XUZs0n7xA4Y

I write about various issues of interest to me that I want to bring to the reader’s attention. While my main work is in Artificial Intelligence and technology, I also cover areas around politics, education, and the future of our children.

In the past year, we have seen a flurry of bold announcements about digital and AI skills in South Africa. Microsoft has committed to training one million South Africans in AI, cybersecurity and related digital skills by 2026. Government departments talk about upskilling tens of thousands of public servants in digital literacy, cyber awareness and AI fluency.

The National School of Government now offers “AI fluency” and Fourth Industrial Revolution courses, while PSETA and NEMISA promote digital skills platforms for officials. It sounds impressive – and in many ways it is. But a nagging question remains: training one million people, for what exactly? If these programmes are not connected to how public jobs are designed, how systems work and how performance is measured, we risk turning a serious national challenge into PowerPoint theatre.

CONTEXT AND BACKGROUND

The basic story is straightforward. South Africa has recognised that it cannot talk about digital government or AI adoption when many public servants still struggle with legacy systems, fragmented data and a lack of confidence around technology. In response, several initiatives have emerged.

Microsoft’s AI skilling initiative, strongly endorsed by the Presidency, aims to provide one million South Africans with training opportunities in AI, machine learning and cybersecurity by 2026. The programme explicitly includes public servants alongside youth and private sector employees. At the same time, the Department of Communications and Digital Technologies has committed to training 30 000 government employees in digital literacy, cybersecurity awareness and AI fluency in a single financial year, through partnerships with NEMISA, the National School of Government, the DPSA and industry players.

The NSG itself has rolled out a suite of open e-learning courses for officials, including modules on AI fluency, cybersecurity and broader 4IR topics. PSETA’s Batho Pele Digital Skills platform and NEMISA’s digital skills programmes add to this ecosystem, giving public servants free access to online courses on everything from basic digital literacy to more advanced topics. On paper, the architecture looks encouraging: multiple institutions, international partners, national strategies and ambitious numeric targets.

INSIGHT AND ANALYSIS

The problem is not the ambition; it is the theory of change. Too often, digital skills initiatives are treated as if training alone can transform a paper-driven, compliance-heavy bureaucracy into a responsive digital state. A two-hour AI fluency video series, or a once-off online course in “digital transformation”, will not magically fix broken workflows, outdated legislation or misaligned incentives.

We should ask some harder questions. Are these courses embedded in career-long learning pathways, or are they stand-alone workshops that officials attend once and forget? Does completing an AI or digital literacy course change anything about how a person’s job is defined, what systems they can access, or how their performance is evaluated? Or is it simply another tick-box on a training report?

There is also a risk of over-promising what AI-specific training can achieve when the basics remain shaky. In many departments, officials still struggle with slow networks, unreliable devices and legacy applications that make even simple tasks painful. Teaching people about generative AI and data analytics while they cannot reliably log into core systems is like offering flying lessons to someone who still has to walk 10 kilometres to get water.

Finally, we need to be honest about measurement. How will we know whether training one million people in AI-related skills has made a difference? Counting enrolments and certificates is easy. Showing that turnaround times improved, error rates dropped, citizen satisfaction increased, or fraud detection got better is much harder – but that is what matters.

IMPLICATIONS

A fair assessment must do two things at once: acknowledge that these skilling initiatives are necessary and valuable, and insist that they are not sufficient. For policymakers, the implication is that digital and AI training must be tightly linked to job redesign, systems reform and organisational change. When a public servant completes a course, there should be a clear path to apply that knowledge: new responsibilities, improved tools, and leadership support to question outdated processes.

For the institutions delivering training, the challenge is to move beyond generic content. Public servants do not need abstract lectures on “the Fourth Industrial Revolution”; they need practical, context-specific guidance on using digital tools to fix grant backlogs, improve clinic scheduling, manage infrastructure maintenance or analyse case data. AI can play a role here too – for example, by providing adaptive learning, practical simulations and role-specific guidance – but only if the curriculum is grounded in real public-sector tasks.

The private sector also has a role. Corporate-funded training should avoid becoming vendor marketing dressed up as skills development. The test should be whether a course leaves a public servant better able to solve real problems for citizens, not just better versed in one company’s product set.

CLOSING TAKEAWAY

Training one million people in digital and AI skills is an inspiring headline, and given our skills crisis, it is a step in the right direction. But if those programmes are not stitched into the fabric of how government actually works, we will end up with exactly what many citizens already fear: certificates on walls, tender reports full of impressive numbers – and very little visible change in frontline services.

The real measure of success will not be the number of officials who can define “artificial intelligence”, but the number who can use digital tools to deliver cleaner data, faster decisions and more humane services. Skills matter enormously, but without redesigned jobs, working systems and courageous leadership, they will remain potential rather than power.

Author Bio: Johan Steyn is a prominent AI thought leader, speaker, and author with a deep understanding of artificial intelligence’s impact on business and society. He is passionate about ethical AI development and its role in shaping a better future. Find out more about Johan’s work at https://www.aiforbusiness.net