Organisations are no longer asking if they should automate, but who survives the transition.
While advanced artificial intelligence is reshaping data-intensive industries from finance, call centres, education and law, tech experts emphasise a critical point: this wave is not designed to eradicate employment, but to redefine it.
As South African firms leverage initiatives like the International AI Readiness Accelerator to develop more capable systems, the urgent question is no longer job preservation, but whether the workforce possesses the skills to work alongside them.
Some sectors have already begun mobilising stakeholders to incorporate AI in their operations. One such sector is hospitality, which stated in its recent annual gathering, The Hospitality Industry Think Tank, that “AI will not replace your job, but someone using AI will.” In hospitality, AI is being framed as a way to free up time for the human touch such as more personalised service, better guest insight, and meaningful interactions that build loyalty.
Seati Moloi, chief executive and founder of Khoi Tech, said sectors most likely to be affected by the new generation of AI tools in South Africa are those that are data-intensive, process-heavy, or dependent on large volumes of repetitive knowledge work.
The sectors likely to be affected include banking, insurance and financial services, call centres, and customer service, professional services such as legal, accounting, consulting, marketing and media, technology and software development, mining, manufacturing and energy, healthcare, education, retail, agriculture, logistics, transport, government and public services.
Moloi clarified that positions of employment being “affected” did not necessarily mean jobs disappearing.
“In many industries, the bigger change will be how people work, which skills are valuable and how quickly organisations can deliver services. The World Economic Forum already identifies AI and machine-learning specialists among South Africa’s fastest-growing roles, while AI and big-data skills are rising in importance across industries,” he said.
Guidelines for working with AI tools
Moloi said in the banking, insurance and financial services, AI is increasingly being applied to fraud detection, credit and risk analysis, compliance, customer service, personalised financial products and internal operations.
In call centres, and customer service AI can answer routine enquiries, summarise calls, retrieve customer information, assist human agents and increasingly complete transactions themselves.
“This is particularly relevant to South Africa because process outsourcing has been an important employment opportunity,” he said.
In professional services such as legal, accounting, consulting, marketing and media, AI can already draft reports, analyse contracts, conduct research, generate marketing material, prepare presentations and summarise large volumes of information.
“The likely outcome is not the disappearance of these professions, but a shift towards professionals using AI to perform routine work while spending more time on judgement, strategy, relationships and complex decision-making. Globally, generative-AI skills are already increasingly appearing in content, marketing and research-related work,” Moloi stated.
He added that the biggest barrier to AI adoption is often not the technology itself, but whether a company’s existing systems, data and infrastructure are ready to support it.
“For many organisations, adopting AI will first require modernising legacy platforms, improving data quality and strengthening security and integration capabilities,” he said.
Moloi said that as AI becomes more capable, people must also become better at checking its work to identify weak assumptions, factual errors, bias, and incomplete reasoning, rather than simply accepting an AI-generated answer because it sounds convincing.
“Employees should understand how to interpret data, recognise trends, question data quality, and distinguish correlation from meaningful insight. They do not necessarily need advanced statistics, but they should understand enough to evaluate the evidence behind an AI recommendation,” he said.
Dr Seani Rananga, a computer science lecturer and an AI systems developer from the University of Pretoria, said global AI tools should not be adopted unchanged; instead, they must be evaluated against South African laws, languages, culture, infrastructure and workplace realities.
“Under POPIA (Protection of Personal Information Act), employees should avoid entering customers’ personal information, confidential records or organisational data into public AI systems unless the organisation has approved the tool and established appropriate safeguards. Organisations should know what data is collected, where it is processed, how long it is retained and who may access it,” Rananga stated.
She added that computer science faculties should work directly with professionals in law, healthcare, education, journalism, finance, public administration, agriculture and the humanities to build a workforce capable of directing AI rather than being displaced by it.
Computer scientists understand how AI systems are designed and evaluated, while industry professionals understand the operational context, risks and consequences of mistakes. Linguists and communities are also essential when systems are developed for African languages because translation alone cannot capture cultural meaning, tone, idiomatic expressions and context.
Professor Nyna Amin, an interim director for Learning and Teaching and an AI governance researcher at the University of KwaZulu-Natal, said that for computer-generated literary and artistic works, the author is the person who undertook the arrangements necessary for the creation of the work.
AI and labour law
She added that South African law is governed by the Copyright Act 98 of 1978 and the Copyright Amendment Bill, which remains in force.
‘Where an employee prompts, curates and edits AI output in the course of employment, Section 21(1)(d) vests ownership in the employer. Two caveats matter. First, copyright requires originality, meaning sufficient human skill and effort. Output generated with a single trivial prompt and no meaningful human shaping may attract no copyright at all, leaving it unprotectable against copying by competitors,” Amin said.
She added that the terms of service of the AI provider govern the contractual position between company and vendor, so these must be read before deployment.
Amin stated that where AI adoption leads to restructuring, the Labour Relations Act applies in full.
She highlighted that automation is a legitimate operational requirement, but Section 189 obliges the employer to engage in a joint consensus-seeking process before any dismissal is contemplated as final.
Tips for the workforce
1. Develop AI literacy. Understand what AI can do, where it fails, and how it may produce convincing but inaccurate information.
2. Strengthen your domain expertise. AI can generate an answer, but professional knowledge is needed to determine whether that answer is appropriate and correct.
3. Learn to manage AI-assisted workflows. Practise breaking tasks into steps, providing clear context, and reviewing outputs systematically.
4. Protect data and verify information. Do not upload sensitive information carelessly, and never treat AI-generated content as automatically factual.
5. Invest in distinctly human capabilities. Critical thinking, creativity, ethical judgement, communication, empathy and contextual understanding become more valuable as routine work is automated.
5. Improve your data and digital literacy. You do not need to become a programmer, but understanding data, automation, AI agents and digital workflows will make it easier to identify where AI can improve your work.
6. Learn to work with AI, not against it. Become comfortable using AI tools for research, analysis, writing, planning, and problem-solving. Employees who can use AI effectively will often be more productive than those who avoid it.
7. Develop critical thinking and verification skills. Do not accept AI outputs at face value. Learn to question assumptions, verify facts and recognise errors, bias or weak reasoning.
gcwalisile.khanyile@nationalmg.co.za