OPINION
The report of Joburg call-centre workers losing jobs amid a shift toward artificial intelligence should change the way African leaders discuss AI skills. Training cannot be treated as a side programme that begins after automation has already removed the work through which young people build experience.
Across the continent, governments and institutions are investing in AI capacity. The African Telecommunications Union and the United Nations have announced collaboration on locally grounded AI skills and digital public infrastructure. UNESCO has also urged technical and vocational institutions to integrate AI through an institution-wide approach that protects human agency, equity and safety.
Those efforts are necessary. But the most important classroom for AI capability may be the workplace itself.
Entry-level jobs in customer service, administration, finance, media, logistics and technology do more than produce immediate output. They teach how organisations operate. A junior employee learns to recognise an unusual case, ask for help, understand a customer, spot an error and balance speed with quality. These are the judgment skills that experienced workers later apply in supervisory and specialist roles.
If AI removes the routine tasks without creating a new way to learn the judgment around them, Africa may gain short-term efficiency while weakening its future leadership pipeline.
Every large AI deployment should therefore include an apprenticeship plan. Employers should identify which capabilities workers currently develop through the affected job and show how new employees will gain them after automation. The plan could combine supervised AI use, rotations, simulations, mentoring and progressively harder assignments.
Training should start with a real workflow. A learner might use an approved tool to draft a customer response, organise a case file or summarise operational data. A qualified reviewer should assess the result, explain what the system missed and require the learner to justify whether the output should be accepted, changed or rejected.
This approach teaches both capability and restraint. It also produces evidence. Employers can measure whether the redesigned workflow saves time after verification, reduces errors, improves service and helps workers advance into more skilled roles. Course attendance and software logins cannot answer those questions.
The transition must also include worker voice. Frontline employees encounter AI failures before executives do. They see when a system misunderstands a local language, ignores social context, gives an unaffordable recommendation or creates correction work for colleagues. Leaders need channels through which employees can report those problems without being branded as opponents of innovation.
African labour markets make this especially urgent. Many countries have young populations, high unemployment and limited formal career ladders. Removing entry points without building replacements will widen the gap between people who already possess experience and those trying to obtain it.
The risk is not limited to job numbers. Imported systems may perform poorly with local languages, informal business practices and public-service realities. An employee who understands the community often supplies context the model lacks. Organisations should record where human intervention was needed and use that evidence to improve local datasets, instructions and review standards. This turns everyday work into a source of African capability instead of treating the workforce as a passive user of foreign technology.
Governments can use procurement and incentives to shape better behaviour. Public contracts for AI-enabled services should require vendors to explain how workers will be trained, how consequential outputs will be reviewed and how incidents will be recorded. Tax incentives and training subsidies should reward verified job transitions rather than the number of certificates issued.
Training providers should also be accountable for employment outcomes. A programme should report how many participants practised on supervised workplace tasks, how many entered relevant roles, how many advanced after six or twelve months and whether women, rural learners and people with disabilities received equal access to the opportunity.
Small and medium-sized businesses need a practical route as well. They can begin with one recurring process, establish a baseline, train a small mixed-experience team and review the results after thirty days. Industry associations can provide shared templates, coaches and secure tools so that smaller employers do not have to invent governance alone.
Africa should also invest in roles that automation creates around the technology: workflow owners, reviewers, data stewards, safety testers, language specialists and trainers. These jobs should not be reserved only for experienced technical professionals. Structured pathways can help customer-service workers, administrators and recent graduates move into them.
The continent does not have to choose between technological progress and worker protection. Good adoption connects the two. AI should take over tasks while people gain better work, stronger judgment and clearer routes into responsibility.
The test of Africa’s AI skills agenda should not be how many people attend a course. It should be whether more people can enter the economy, learn through real work and build careers in which technology expands rather than closes their opportunities.
—
Dr. Gleb Tsipursky, a behavioral scientist and author of The Psychology of AI Adoption at Work: From Resistance to Results (Georgetown University Press, 2026).
Author contact: gleb@disasteravoidanceexperts.com
Disclaimer: The views expressed in this Opinion article are those of the author and do not necessarily reflect the editorial position of Towncrier Africa.
About the author: Dr. Gleb Tsipursky is a behavioural scientist and author of The Psychology of AI Adoption at Work: From Resistance to Results (Georgetown University Press, 2026).
Discover more from Towncrier Africa
Subscribe to get the latest posts sent to your email.
