Table Of Contents

Beyond the North-South Fork on the Road to AI Governance:

An Action Plan for Democratic & Distributive Integrity*

* Acknowledging that the categories of South and North are not watertight, this paper argues for situating geopolitical and geo-economic power within the history of post-colonial development.


Justice in the AI Economy

Fair Value Distribution & Implications for Development

AI is to our digital epoch what electricity was to the industrial revolution: a paradigm-shifting, general-purpose technology whose diffusion brings an exponential increase in productivity. Such increase derives from the augmentation of fixed capital and human capabilities in the production process, labour substitution, and product and service innovation (Bughin et al., 2018; Zuboff 2018). AI is estimated to add anywhere between USD 13 trillion and USD 15.7 trillion to global economic output by 2030 (Rao & Verweij, 2017). As the United Nations Conference on Trade and Development’s (UNCTAD) 2021 Digital Economy Report observes, business models revolving around AI cannot exist without control over the data that feeds such models (UNCTAD, 2021). The generation of “intelligence premium” (Gurumurthy et al., 2019) is predicated on the ceaseless capture of social data7. This explains why the first-mover digital platforms from the US and China that control huge data enclosures are also leading investments and research in AI8.

If global AI adoption continues along the same trajectory, it might widen performance gaps, not just at the firm level and the individual worker level, but also the country level. Front-runner AI companies are likely to benefit disproportionately and may double their returns by 2030 while companies that delay adoption will be left far behind (Bughin et al., 2018). Similarly, at the worker level, demand for jobs and wages may grow for a few knowledge workers with digital and cognitive skills and with expertise in tasks that are hard to automate, but will shrink for the majority performing repetitive and low digital skill jobs (Acemoglu et al., 2020). The US and China dominate the entire global AI economy: the two countries account for over 94 percent of all funding of AI start-ups in the past five years, 70 percent of the world’s top AI researchers (UNCTAD, 2021) and 90 percent of the market capitalisation value of the world’s 70 largest digital platform companies that control a significant proportion of cross-border data flows on the Internet (UNCTAD, 2019). American and Chinese participants are also better represented in the industry bodies that develop standards, creating long-term dependency on basic technical protocols for the whole world.

With big data drawn from the Internet of Things becoming crucial, the EU, South Korea, and Japan, with their strong manufacturing base, associated computing power, and human resource capabilities, stand a very good chance of catching up (UNCTAD, 2021). The winners may well dominate the coming decades geo-economically and geo-politically (Feijóo, et al., 2020).

The acquisition of effective domestic AI capabilities depends upon three factors: big data, computing power, and the work of prominent AI researchers and engineers. Unfortunately, developing countries, disadvantaged both by the adverse terms of their integration into the Internet economy of user-generated data flows and limited industrial capacity to shift to smart manufacturing, are at high risk of being relegated in perpetuity to the low value parts of the AI economy. As currently configured, the AI race threatens to leave sub-Saharan Africa and most developing countries behind (UNCTAD, 2021), with an unprecedented concentration of wealth in the hands of a few companies in China and the United States. The competitive advantage in their ‘cheap labour’ that developing countries historically enjoyed may thus be rendered completely irrelevant (Lee, 2018).

On a fine-grained level, the consolidation of data ownership in the hands of big technology multinationals feeds into local inequalities in countries of the Global South where they operate. This asymmetry in data ownership represents a barrier to entry for smaller homegrown start-ups and feeds into market concentration in contexts where local legislative infrastructure is weak and laws on competition and data protection, if present, are still nascent (Rizk, 2019). This exacerbates inequality and results in further exclusion for the less fortunate in countries of the Global South.

The lack of a globally accepted economic resource governance regime for data aids economic concentration and deepening of inequalities in the AI paradigm9. The rules for cross border data flows in the global economy are determined by a few powerful countries whose corporations enclose data from far and wide as trade secrets (James, 2021), asserting de facto ownership rights over these holdings (Fia, 2021). In this intelligence economy, countries and communities of the Global South lacking in data processing and AI capabilities face a dangerous and untenable paradox. Not only must they relinquish any claims to their own data now locked up in AI systems of transnational capital, but they also have no means to legitimately derive a fair share of benefits generated therein. This results in gross economic unfairness in the global digital economy. Algorithmic coloniality is thus naturalised (Gurumurthy & Chami, 2021).

  1. Mass digitisation, which expanded with the Internet in the 1990s and escalated with data centres in the 2000s, has made available vast resources of data. A regime of knowledge extraction – built on Big Data – gradually employed efficient algorithms to extract ‘intelligence’ by capturing these open sources of data, mainly for the purpose of predicting consumer behaviour and selling ads. The knowledge economy has morphed into a novel form of capitalism in which unilateral control over data-based intelligence is the source of profit.
  2. As UNCTAD (2021) observes, between 2016–2021, there were 308 merger and acquisition (M&A) deals worth $28.4 billion in the AI start-ups segment. The top five companies in the world, by number of acquired AI start-ups in the same period, were the Big Tech companies from the United States, followed by Baidu (sixth) and Tencent (eighth) from China. Apple led this ranking, followed by Google and Microsoft.
  3. That said, there may be other rights-based regimes we need to establish before we start institutionalising a regime for data as an economic resource.