Beyond job loss: AI as Africa's productivity co-pilot | B&FT

TEXT START: As artificial intelligence begins to reshape the world of production, there remains a question of what role AI will play in the continent of Africa.

The Dissection

The article reframes AI from labor-displacing technology into a productivity aid. It accepts that African workers are underproductive, then presents AI as a co-pilot that raises output without asking who owns the systems, captures the surplus, or decides whether workers remain necessary. It converts a distributional conflict into a technical adoption problem: implement AI correctly and everyone benefits.

That contains a partial truth. AI can initially raise individual output, improve training, reduce errors, and expand the capabilities of undercapitalized firms. But co-pilot is a phase description, not a stable endpoint.

The Core Fallacy

The article confuses higher productivity per worker with greater demand for workers. Under P1, when AI allows one worker to perform the work of several, competitive pressure rewards firms that reduce headcount. Productivity gains benefit producers; they do not automatically preserve wage-based participation.

Africa’s labor surplus makes the contradiction sharper. AI can make workers more productive while making fewer workers necessary. If the systems, data, compute, and capital are externally owned, much of the value can leave the continent while local labor absorbs the displacement. The co-pilot eventually becomes the autonomous workflow, and the worker becomes an avoidable interface.

P2 prevents stable human-only economic zones at scale. P3 follows: the majority lose access to economically necessary labor even as aggregate output rises.

Hidden Assumptions

• Demand will expand quickly enough to absorb workers made redundant by productivity gains.
• Every automated task will generate enough complementary human work to replace what it destroys.
• Firms will retain workers after AI makes their roles cheaper to eliminate.
• African states, firms, and workers will own or control the AI infrastructure and resulting surplus.
• AI will remain a co-pilot rather than becoming a cheaper competitor.
• Better skills and education can outrun the rate of cognitive automation.
• Economic growth will automatically translate into broad income and productive participation.

These assumptions are not demonstrated. They are smuggled in to make technological displacement look like development.

Social Function

Primary classification: transition management and ideological anesthetic. Secondary classifications: partial truth, elite self-exoneration, and prestige signaling.

The article gives governments, employers, and technocrats a reassuring vocabulary for deploying AI without confronting ownership or distribution. It describes labor compression as empowerment and postpones the question of who receives claims on the additional output. The framing may be sincere, but sincerity does not alter the mechanism. It is a polished lullaby for the lag period before replacement becomes visible.

The Verdict

The article correctly identifies AI’s immediate capacity to raise African productivity and incorrectly treats that capacity as a defense of mass employment. It mistakes the first stage of obsolescence for the endpoint.

Under the Discontinuity Thesis, AI will be introduced as a co-pilot because institutions and firms initially still require human interfaces. Competition then rewards organizations that turn those co-pilots into autonomous systems and shed labor. Africa’s decisive question is not whether workers can become more productive. It is who owns the productive machinery, controls energy, logistics, maintenance, and compute, and distributes the resulting surplus.

Without ownership or indispensability, AI productivity is not a rescue. It is a more efficient route to exclusion.