Whether one views AI as a bubble or a boom, it must eventually end. If it is a bubble, AI may be sustained as improbably long as cryptocurrency, but it will inevitably subside. However, if AI is a burgeoning general technology, it will eventually become embedded in various other products and services. At that point, AI will no longer draw the same levels of investment and public scrutiny that it currently does. One question remains invisible in the formulation of AI policy across the African continent: What will the legacy of AI be, and specifically, what infrastructure will remain after AI?
The shaping of AI's contribution to the future, through policy, implementation, and investment—whether aligned with national processes or not—seems curiously elided in the current AI debate.
Lessons from South Africa's Minerals Revolution
In contemplating the end of AI in Africa, it is useful to reflect on the minerals revolution in Southern Africa that began in the 1860s and reshaped the region over the next four decades. The politics and economics of the region were completely reshaped, especially through racial and class stratification, that have not yet been undone.
For purposes of informing AI policy debate, the stark contrast between the legacies of Kimberley and Johannesburg, shaped by diamond and gold respectively, repay attention.
Diamonds were first discovered in the 1860s, rapidly transforming the Griqua territories in which they were found. After negotiations between the Cape Colony, Zuid-Afrikaanse Republic (ZAR) , the Orange Free State, and the Griqua leader, the territory was incorporated into the Cape Colony.
The site of the first diamond mine, Kimberley, grew quickly, and by November 1885, railways had reached it, deep in Southern Africa's interior. By that time, the city had already had electric street lighting since 2 September 1882. Kimberley was the first city in the Southern Hemisphere and the second in the world, after Philadelphia, to use electrical street lighting.
Initially, small syndicates of individual diggers worked the diamond fields, but they were soon replaced by larger companies, and ultimately the monopoly of De Beers. The colonial powers did not impede this monopolization, and the chief monopolist, Cecil John Rhodes, became Governor of the Cape Colony. Meanwhile, gold was discovered on the Witwatersrand in the independent Transvaal (ZAR). Mining capital, frustrated in its attempts to use the state to structure the political economy of the ZAR precipitated the Second Anglo-Boer War, leading to British control of the area.
Johannesburg, founded in 1886, grew rapidly, building an internal railway by 1888 and connecting by rail to the Cape by 1892. Like diamonds, gold quickly came under monopoly control. The investment of capital into diamonds and gold drove the political economy of Southern Africa, with mining capital seeking to leverage the state to extract labour from self-sufficient African farmers.
Questions of Infrastructure (i)
This brief account of the minerals revolution raises the question: what kinds of infrastructure resulted from this period?
The railways and harbours built to ship minerals to the colonial powers were significant, but equally important were the knowledge infrastructures, such as the blasting certificate, a qualification that allowed a working class white miner to earn a higher wage in the mines. When mining capital sought to extend this certificate to Black miners in an effort to lower costs, it led to an armed uprising in 1922, so important was this infrastructure.
The subsequent histories of Kimberley and Johannesburg diverged sharply. Kimberley, though remaining an important town which eventually became the capital of Northern Cape province in 1996 remains a relatively small city, its economy reliant on the remnants of mining and dryland agriculture. Johannesburg, by contrast, grew into a megalopolis and the financial centre of Southern Africa, even as mining operations shifted further afield. Though always controlled indirectly from London, mining capital formally relocated when Anglo-American was listed on the London Stock Exchange in 1999.
Questions of Infrastructure (ii)
The vastly different infrastructural legacies left by the minerals revolution in Southern Africa prompt a crucial question for the future of AI in Africa:
What kinds of infrastructure will AI leave behind, and how will it affect different regions of the continent?
What kind of foundation will AI, shaped by AI policy lay for future innovation, economic participation, and governance, but what exactly will these legacies be?
Corey Doctorow, describes AI as a bubble, but does not dismiss the question of what will succeed it. He points out: “Tech bubbles come in two varieties: The ones that leave something behind, and the ones that leave nothing behind. Sometimes, it can be hard to guess what kind of bubble you’re living through until it pops and you find out the hard way.”
Cory makes it clear that the question of what comes after AI is not only being elided in Africa, but in the global North:
Our policymakers are putting a lot of energy into thinking about what they’ll do if the AI bubble doesn’t pop – wrangling about ‘AI ethics’ and ‘AI safety.’ But – as with all the previous tech bubbles – very few people are talking about what we’ll be able to salvage when the bubble is over.
We should ask what kind of legacy AI might leave. Will it produce a generation of people trained in technical skills that can be applied to other tasks? What physical infrastructure will remain, and what regulatory and institutional frameworks will emerge in response to AI policy?
In this context, it is essential to also ask what institutions created, or changed in response to AI policy will survive and acquire momentum of their own? What will be the long-term effects of dedicating resources to AI rather than to other developmental priorities? And what will be salvageable from AI investments when the technology no longer commands the same level of attention and resources?
AI Infrastructure Investments
AI policy must consider how to "fail well”. If AI Large Language Models prove to be little more than lossy compression algorithms the investments made by private and public actors across Africa may still yield useful outcomes. Sustainable data centres that use renewable energy and sustainable water sources, for example, could remain valuable assets, as AI is merely the latest data-dependent technology driving the need for such infrastructure. These centres could provide a foundation for future technologies reliant on data analytics.
Moreover, broad, flexible laws designed to regulate autonomous processes could provide enduring frameworks for future technologies. But laws written narrowly for AI defined as no more than the current iterations of large language models will become outdated as soon as the technologies that prompted them.
In due course, artificial intelligence will no longer occupy the central locus of public and policy discourse. This eventuality may arise either from AI's successful integration into quotidian existence - thereby becoming an embedded aspect of routine life - or from its failure, leaving in its wake a set of functional yet constrained instruments.
Regardless of the trajectory, the fundamental question persists: What will follow the era of AI?
Comments
Post a Comment