Data, AI and society

  • Blog

What do we mean by data for development?

Without addressing the significant epistemological and ideological challenges associated with the concept of development, over the last five years, Research ICT Africa’s work has shifted from examining the effects of digitalisation on economy and society to the policy and governance challenges of ‘datafication’ for developing countries.

The initial effects of digital mobile technologies brought basic voice communications to billions of people across the Global South. Although access and use were highly uneven, it was only once broadband was introduced and the shift from voice to data services took place that the digital inequality paradox became truly evident as a wicked policy problem. Arguably, the biggest planetary policy challenge after climate-change, this refers to the fact that as we inevitably connect more people, digital inequality increases.

From a policy perspective, it is not a new problem and has been referred to for years as the digital divide in the context of voice services. However, In a data environment, unlike in voice, this is the case regardless of whether people are online or not. Inequality also exists between those barely connected, i.e., those going online for a few minutes to passively consume services with small data bundles they can barely afford, and those people able to optimise their use of the Internet to improve their wellbeing. Those able to do so can reduce the transaction costs associated with consumption of goods and services; or even deploy digital technologies online for productive purposes. In sufficient numbers, active consumers and productive users contribute to the prosperity of nations, compounding the uneven development of nations and indeed globalisation.

From the modelling of the nationally representative demand-side 2018 After Access surveys of 20 countries in the Global South, we know that factors behind the intersectional nature of digital inequality are education and the associated factor of income. Digital disparities in gender, for example, broadly reflect the fact that women are concentrated amongst the most marginalised – poor, rural, unemeployed. Countries with higher GDP per capita (higher levels of education and formal employement) tend to have digital gender parity. From the demand side, redressing these challenges in a data environment then represents a classical human development challenge – one not simply addressed by ‘connecting the last billion’, though this is a precondition for development.

Unless concerted efforts to readdress them are made from a political economy perspective, the overlaying of processes and technologies of datafication on unevenly evolving processes of digitalisation – particularly as it occurs through global corporations and private monopoly digital platforms – are likely to amplify underlying structural inequalities.

We argue that this will require reframing policy and governance to deal with the global nature of digitalisation and datafication, if we are seeking public interest outcomes.


How can data be governed as a public good?

There are two related aspects of datafication that I want to highlight in this regard. The one relates to the ongoing absence of data or access to it in Africa for evidence based policy and governance of data, despite the massive amounts of data being generated daily. We believe this requires ensuring data is governed as a public good in the classical economics sense – non-rivalrous and non-exclusive – in the same way as national statistics are.

The emerging data required for evidence based policy and regulation challenges relate directly to these dynamic processes of digtialisation and datafication. Given the governance challenges associated with increased data extraction from marginalised citizens, there is also a need to gather new forms of data and indicators across the digital economy. There is much potential for digital sensors and big data to provide some information at a much lower cost and substitute for certain kinds of high cost public surveys. However, they are unable to provide the disaggregated data required to identify the exact points of policy intervention to address new aspects of marginalisation resulting from datafication. Only in-depth demand-side surveys are able to do this in pre-paid mobile markets and conditions where people are entirely marginalised from services.

Integrated supply, demand and big data gathering systems are required to ensure data is governed as a public good. With data now the most valuable resource in the economy, far more innovative governance of it is necessary to ensure public interest outcomes. For example, regulated entities could be required to provide their historical data in manipulatable form, rather than ineffectual universal service levies.

Global platforms facing pushback on their unaccountable data extraction could be required to make historical data available for public use, as part of global governance initiatives. In the context of open data policies that safeguard people’s rights to privacy and anonymity, this would enable the free flow of information required for more effective planning by government and service delivery entities, increase the uptake of online services as well as create opportunities for entrepreneurialism and innovation.


Governance of global public goods

This highlights the challenge of developing regulatory and governance frameworks where generators and users of data are not regulated entities. Platforms may have no physical presence in the jurisdiction in which they operate; and citizens may be data subjects without ever being connected to a network. This is especially the case where private entities are delivering public goods in unregulated data environments such as with biometric identification for civil registration or refugee registration. The supply side data or metadata may not be available to authorities, nor may states have the capacity to protect public data from commercial use. Issues of privacy and data protection for beneficiaries of digital identification need to be understood in addition to the marginalisation of the most vulnerable from these systems, as highlighted in the lessons from Aadhar the Indian biometric system.


Research approach to digitalisation and datafication

These issues form part of RIA’s new research agenda for 2020 -2023. We will consolidate the ‘rights and risks’ framework to mitigate the harms associated with the digital inequality developed in the first round. This highlighted the vulnerability of large numbers of people in Africa coming online for the first time to the abuse of their rights (of which they were often not even aware). This round will move beyond researching degrees of compliance at the national level to global norms of cybersecurity and data protection to exploring what is required to realise data justice.

RIA examines this phenomenon from a political economy perspective coupled with evolving commons theory and need for demand side valuation for a more equitable allocation of resources than the commercial supply side only assessment that has characterised market reforms for the past two decades (Frischmann, 2004, Hess, 2000; Ostrom, 2009). Despite the global nature of most digital and data governance challenges, policy aimed at reducing digital inequalities must still be resolved at a national level. The Internet, data, cybersecurity and so on, only emerge as global public goods to the extent that countries produce them at the national level (Kaul et al. 1999).

It will also extend its focus to the effects of digitalisation and datafication systemically, not only on the individual. While citizens may not be directly affected as a result of not being connected, the processes of digitalisation and datafication may result in systemic improvements to efficiency, better information flows and lower costs for transactions.

In this regard, RIA will extend its approach of developing alternative strategies to those developed in mature markets with strong institutional endowments to realise public interest outcomes. Specifically, it will be exploring how advanced technologies that are being developed and deployed on the continent in the private sector might be effectively used in the public sector to address public delivery challenges. It will also explore how advanced technologies of AI, drones and satellite imagery, deployed by private firms, could be extended in the informal sector or by subsistence farmers existing alongside them, through public-private interplays, co-operatives or other commons initiatives.


References

Best, J., & Gheciu, A. (2014). The return of the public in global governance: Cambridge University Press.

Frischmann, B. M. (2012). Infrastructure: The social value of shared resources: Oxford University Press.

Gillwald, A., & Van der Spuy, A. (2019). The Governance of Global Digital Public Goods: Not Just a Crisis for Africa. Paper presented at the GigaNet, Berlin.

Hess, C., & Ostrom, E. (2003). Ideas, artifacts, and facilities: information as a common-pool resource. Law and contemporary problems, 66(1/2), 111-145.

Kaul, I., Grunberg, I., & Stern, M. (Eds.) (1999). Global Public Goods: International Cooperation in the 21st Century. : Oxford University Press.

Ostrom, E. (2009). Building trust to solve commons dilemmas: Taking small steps to test an evolving theory of collective action. In Games, groups, and the global good (pp. 207-228): Springer.

Related