Disruption Without Dividend

The World Bank's headline says developing economies face little automation risk, and the background paper it commissioned says the disruption will arrive before the benefits do.

Sign up for my Substack daily AI newsletter here.

See my AI Training course portfolio for corporate Business Leaders here.

Article link: https://johanosteyn.substack.com/p/disruption-without-dividend?r=73gqa&utm_campaign=post&utm_medium=web&showWelcomeOnShare=true

Follow me on LinkedIn: https://www.linkedin.com/in/johanosteyn/

I went looking to correct myself. I have argued for some time that artificial intelligence will cost more jobs than it creates, and a major new report appeared to contradict that. Reading it properly produced a different result, and the reason lies in the difference between what an institution announces and what its researchers found.

CONTEXT AND BACKGROUND

The World Bank released its World Development Report 2026, subtitled The Promise of Artificial Intelligence, on 4 August. Its central finding on employment is reassuring. Jobs in high-income countries are more than three times as likely to face automation risk as those in low- and middle-income countries, where 4.5 percent of existing jobs are exposed against 14.2 percent in wealthier economies, while 16.2 percent of jobs in developing economies could see productivity meaningfully boosted, close to the 18.7 percent expected in rich ones. The report concludes that the greatest promise for developing countries lies not in replacing workers but in amplifying what they can do, with Chief Economist Indermit Gill describing AI as a lifeline that developing economies should seize.

Underneath that report sits a joint working paper by the International Labour Organization and the World Bank, prepared as a background study for it, covering 135 countries and roughly two thirds of global employment. It is titled “Disruption without dividend?” and its conclusion is that in developing economies disruption may materialise faster than productivity gains.

INSIGHT AND ANALYSIS

The background paper explains why a low aggregate figure is the wrong comfort. Workers in jobs vulnerable to automation are often already online, even in low-income settings, so losses can arrive quickly. Those jobs represent relatively higher-quality work in poorer countries, including clerical and administrative positions that have historically offered a pathway to decent work, particularly for women and young workers, and the researchers warn that automation could close off those pathways. Meanwhile many workers in roles with genuine potential for productivity gains lack reliable internet, so the gains cannot reach them. Connectivity is acting as an asymmetric filter. It is sufficient to deliver the harm and insufficient to deliver the benefit, and the same infrastructure gap that makes a rural clinic or a smallholder farm unreachable by useful tools leaves an urban call centre entirely reachable by the ones that replace it. The paper adds that jobs sharing a title across countries involve different tasks, with workers in lower-income economies performing fewer non-routine analytical tasks and more routine or manual work, which further narrows the scope for augmentation.

I have previously written about this in a column on automation and the end of entry-level work, where I argued that the first rung of the career ladder is vanishing, that entry-level roles once served as the bridge between study and skill, and that without early experience young professionals risk becoming a lost middle, too qualified for manual labour and too inexperienced for skilled positions.

IMPLICATIONS

The aggregate hides the arithmetic that matters. Research by Caribou and Genesis Analytics for the Mastercard Foundation found that 40 percent of tasks in Africa’s business process outsourcing and IT-enabled services sector could be automated by 2030, with only 10 percent fully resilient. Entry-level jobs make up 68 percent of that workforce and more than half their tasks are automatable. Customer experience roles account for 44 percent of sector employment with half their tasks at risk, nearly two thirds of junior finance and accounting tasks are exposed, and tasks performed by women are on average 10 percent more susceptible than those performed by men. That sector has been one of the few reliable routes into the formal economy for young South Africans without connections. Our national exposure figure is low for the same reason it is low across the region, being that so much of our work is informal, manual and undigitised. Safety through underdevelopment is not safety. It is a description of the problem.

CLOSING TAKEAWAY

None of this makes the World Development Report wrong. Its framework is sensible and its warning about electricity, connectivity and skills is correct. The lesson is narrower and it concerns how we read. An institution’s press release and the research it commissioned can point in different directions, and the difference is not conspiracy, it is emphasis. One document leads with promise because promise is what a development bank exists to offer. The other asks whether the dividend will arrive at all. The question mark in that title was placed there by economists working for the same institution, and for a country with our unemployment rate, that question is the one that matters.

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