LIONEL ADENDORF | How South Africa’s AI-deficient education reforms betray its students

Between 2021 and 2025, China discontinued or suspended roughly 12,200 undergraduate programmes and introduced about 10,200 new ones, redirecting higher education towards AI, robotics, semiconductors, intelligent finance and digital agriculture, and adding AI education itself as a major in 2025. Officials called the discontinued programmes “obsolete”.

It is an extraordinary and uncomfortable intervention. Humanities, arts and language programmes were among those cut. But we would be foolish to ignore the question China has forced itself to answer: what should a university teach when technology is changing the nature of knowledge and work?

China is using university education as part of industrial and economic policy.

South Africa has just implemented changes of a comparable scale in its vocational training system.

In June, higher education minister Buti Manamela confirmed that 1,475 pre-2009 qualifications, including technical education certificates and diplomas, the backbone of the technical & vocational education (TVET) colleges, had reached the end of their registration period. Of these, 630 received extensions, and the rest are being retired or replaced by occupational qualifications that combine classroom learning and workplace experience.

Read the minister’s explanation, and one word is conspicuously absent: AI

Read the minister’s explanation, and one word is conspicuously absent: AI. The stated reasons for this reform are modernising post-school qualifications, aligning training with workplace needs, and improving graduate employability. All worthy goals, but nowhere is AI named as the reason these qualifications needed to change. China asked what knowledge and capabilities its economy would require in an AI-shaped future; South Africa asked how to bring old qualifications up to date.

China reviewed and restructured undergraduate programmes across its university system. South Africa is undertaking an AI-absent reform within its vocational system while largely leaving university programme structures untouched. Treating these exercises as equivalent understates both deficiencies.

Changing TVET qualifications while retaining outdated university programmes cannot constitute a national AI response.

Indeed, reforming TVET should make university reform more urgent. TVET colleges, universities of technology and traditional universities are part of the same education and skills pipeline. If occupational roles are changing because of AI, automation and data, the professional and academic qualifications above them must also change.

The broader evidence confirms where the gap lies. A government-linked AI Maturity Assessment for South Africa, published in 2025 by GIZ and the University of the Western Cape, finds that South Africa is in the 0%–5% range for AI-related university courses taught in English around the world.

Much of South Africa’s university response remains peripheral: research centres, institutional rules for generative AI, specialised postgraduate offerings, short courses and occasional new modules. These are valuable, but they do not constitute a systematic review of what every graduate must know in an AI-shaped economy.

AI is not creating only a small set of specialist technology jobs; it is also changing the work of artisans and technicians as much as that of accountants and engineers. Is every occupational field being redesigned for work already shaped by AI and automation? If these capabilities are absent from qualifications at any level, we certify people for occupations whose technologies are already changing around them.

AI-shaped societies will need ethics, history, law and philosophy

South Africa should not copy China uncritically. China’s cuts have fallen heavily on the humanities, arts and language programmes. Universities cannot be reduced to factories producing workers for industrial policy.

Our literacy crisis makes language and critical reasoning even more important. AI-shaped societies will need ethics, history, law and philosophy. But those disciplines must also help students understand technology, power, truth, human agency, and dignity in an automated society.

The department of higher education & training, the Council on Higher Education, the Quality Council for Trades & Occupations, the South African Qualifications Authority, universities and TVET colleges should conduct a single national audit of AI-era qualifications across both tiers. Every qualification should be evaluated against how AI, automation and data are transforming its discipline.

New and redesigned qualifications should include an explicit AI target, prioritising robotics, cybersecurity, advanced manufacturing and digital agriculture. Industry must help design curricula.

China is redesigning education at every level around the economy it intends to build. South Africa has demonstrated the institutional capacity to implement reform; but our problem is ambition without alignment. We need qualifications, from TVET colleges to university lecture halls, designed for the future that is already here.

The AI age will not respect the boundary between vocational and university education. Our response cannot stop at that boundary either.

• Adendorf is chair of the board of the National Electronic Media Institute of South Africa, former chair of the Western Cape Community Education & Training College Council, and MD of reputegrity.ai, an AI ethics and governance consultancy