Research ICT Africa’s Submission to the Global Digital Compact

What would a Just Global Digital Compact for Africa look like?

Preface

This submission is based on a public consultation of an earlier discussion paper that draws on the research and evidence base that Research ICT Africa (RIA) has built over two decades to support the
development of contextualised policy and regulatory strategies, as alternatives to those that have
failed to redress digital inequality and data injustice in Africa. The consultation took the form of a call
for comments on the discussion paper, either through email or an online form on the RIA website.
We then held a consultative webinar to engage a multistakeholder audience on the content. A
French version of the webinar presentation was posted for comment. RIA then revised the discussion
document into this submission. This is intended to serve as an evidence base and proposed framing
for the online submission made in response to the UN Secretary General’s call for input into a Global
Digital Compact.
The overarching question we asked in our public consultation, and which informs our research
agenda is: what policies and forms of governance are required to realise global digital public goods
at the national level, to redress digital inequality, harness the potential of new technologies for social
and economic development, improve public sector efficiency and delivery and create public value?
This submission was prepared by: Alison Gillwald, Andrew Patridge, Andrew Rens, Abdiaziz Ahmed,
Araba Sey, Diana Nyakundi, Guy Berger, Liz Orembo, Jackie Okello, Naila Govan Vassen, Rachael
Adams, Roland Banya, Sandra Makumbirofa, Samantha Msipa, Senka Hazdic with contributions from
Steve Song.
We would also like to thank all RIA staff past and present who have built the repository of knowledge
on which we draw, together with our partners across Africa and collaborators across the world.
The authors would like to thank the participants in its online consultations and the donors that made
possible the independent research and the RIA’s contribution to global regional and national policy
processes such as these, particularly current donors IDRC, SIDA, the Bill and Melinda Gates
Foundation and GIZ. The views expressed are those of Research ICT Africa and do not represent or
reflect those of the donors.
For further information contact: Alison Gillwald agillwald@researchictafrica.net or Liz Orembo
lorembo@researchictafrica.net

1. Introduction

Research ICT Africa (RIA) welcomes this call for submissions for stakeholders to contribute to the
development of a Global Digital Compact (GDC). We are an African digital policy, regulation and
governance think tank that has operated for two decades to fill a strategic gap in the development of
a sustainable information society and digital economy. Through rigorous and relevant research and
analysis, RIA seeks to build an African knowledge base in support of digital equality and data justice
and to monitor and review developments on the continent. Our input into the GDC draws on the
evidence base that we have built over two decades of digital policy research, and which supports the
development of contextualised alternative policy and regulatory strategies to those that have thus
far failed to redress digital inequality and intensifying data injustice in Africa.
The digitalisation of the economy and society and the datafication of almost all activities of the
connected has promoted economic efficiency, increased access to the means of production and
communication, and driven innovation. These developments have been extremely uneven however,
both between and within countries, with most the world’s population still marginalised from access,
or from productively using the Internet to enhance their well-being or to improve their livelihoods.
Moreover, inequality is being exacerbated by the layering of advanced technologies such as machine
learning and Artificial Intelligence (AI) over existing intersectional digital inequalities – which already
reflect underlying structural inequalities. These intensifying global processes of digitalisation and
datafication are simultaneously accompanied by a plethora of individual and (particularly poorly
understood and defined) collective risks that, unmitigated, could result in widespread harms to
human rights, including to sustainable development and democracy.
UN Secretary-General Antonio Guterres identified this when he described digitalisation as one of two
seismic shifts that will shape the 21st Century, the other being climate change. He cautioned that
both would widen inequalities even further unless urgently addressed on a planetary scale.
The COVID-19 pandemic has further highlighted the compounding effects of digital inequality on
underlying structural inequality. The uneven capabilities of individuals, formal and informal firms,
and states to mitigate the public health and economic risks associated with lockdowns through the
digital substitution of access to work, schooling, food-sourcing, public services and social relief
programmes have highlighted the centrality of digital inclusion. This bleak scenario has been
compounded by high incidences of geographic, gender, race and ethnic marginalisation and
discrimination, particularly among refugees, who fall out of even the most basic safety nets where
they exist.
This unevenness, marginalisation and exclusion applies not only to economic and social
participation, global competitiveness or the geopolitical positioning of states but also to exercising
effective citizenship. Rather than fostering political inclusion, increased digitalisation is
accompanied by a sense of democratic erosion, disinformation and disorder in an increasingly
digitalised public sphere.

To arrest these negative trends and the lack of progress being made towards achieving the 2030
Sustainable Development Goals, the UN Secretary General has called for a renewal of the social
contract, anchored in human rights and gender equality, to rebuild trust and social cohesion
that people need to see reflected in their daily lives. Highlighting the centrality of digital
inclusion in contemporary society, the Common Agenda, which was borne out of a declaration
by the General Assembly during its commemoration of 75th Anniversary in 2021, calls for a
Global Digital Compact that “should also include updated governance arrangements to deliver
better public goods and usher in a new era of universal social protection, health coverage, education,
skills, decent work and housing, as well as universal access to the Internet by 2030 as a basic human
right so all citizens have a say in envisioning their countries’ futures.”
The realities of underdevelopment in the Global South, combined with the speed of digitalisation,
underscore the need to innovate policy formulation, regulation and global governance to propose
solutions to the wicked policy problem of the digital inequality paradox. The paradox lies in the fact
that as more people are connected to the Internet – which has the potential to enable a new social
compact, support economic and political state formation, and create public data for policy and
planning, job creation, resource mobilisation and redistribution – this, in fact, amplifies structural
inequalities in developing countries as well as lends itself to disinformation operations that
disadvantage already vulnerable people. This has made it critical to find more inclusive and
equitable ways of bringing people online, ensuring the capabilities not only to use general purpose
technologies such as the Internet for personal use but also to deploy them to enhance their lives,
produce value and contribute to the prosperity of nations and an information environment
conducive to democracy and sustainable development.

With its mission of accelerating digital equality and data justice in Africa, RIA has developed a
comprehensive submission in response to the UN Secretary General’s call for inputs on the GDC
as part of the Common Agenda. The overarching question the submission seeks to address is what
policies and forms of governance are required to realise global digital public goods at the national
level, to redress digital inequality, harness the potential of new technologies for social and economic
development, improve public sector efficiency and delivery and create public value.
What is needed to protect citizens from harms associated with being data subjects of big data
collection and analytics and to mitigate the risks of ever-more pervasive extractive monopoly
platforms that are nudging, shifting and modifying our behaviours? This submission contends that
non-siloed, transversal digital and data policy recognises the role of digital public goods as central to
contemporary forms of democratic participation and as key inputs and enablers of economic
transformation, together with human development strategies and rights preserving regulatory
arrangements to redress intersectional inequality and foster integrity in the information
environment. Such a policy, under the right conditions and which acknowledges the political
economy of developing countries, will be essential to post-pandemic economic reconstruction and
the building of more democratic, inclusive and equitable social compacts. The submission proceeds
by identifying the wicked policy problems arising from digital inequality and data injustice, through
an intersectional inequality lens and from an African perspective. It proposes for the GDC a global
digital public goods framing for the global governance of the intensifying process of digitalisation
and datafication and in doing so it surfaces seven critical areas that could contribute significantly to
more equitable and just digital policy outcomes:
▪ public data for evidence-based policy;
▪ connectivity, access, pricing, quality;
▪ e-trade, labour, taxation and social protection;
▪ data governance and data justice;
▪ disinformation and the information disorder; and
▪ harnessing AI for Africa (opportunities and harms).

2.Policy problem

Intensifying globalisation has been both driven and produced by dramatic technological innovations
and giant corporations that have resulted in massive efficiencies and opportunities through
digitalisation and datafication. However, this has been highly uneven with the majority of the world
population far removed from these benefits, but still vulnerable to the harms.

These developments have been ushered in by global oligopoly platforms operating new forms of
extractive capitalism, with platforms amassing large amounts of data on their users, which is used to
both shape their online behaviour and sold to advertisers for targeted advertising. At the same time,
the platforms’ architectures have lent themselves to exploitation by actors whose business practices
infringe human rights, while platform spending patterns have neglected the importance of
monitoring, moderation and risk assessments in the Global South. These global trends have also
enabled the emergence of a raft of disruptive digital and ‘virtual’ low-cost substitutes for physical
services. On the one hand, they have offered new livelihood opportunities, such as Uber, AirBnB,
AmazonTurk and Fiverr, but on the other, they have operated outside of traditional market
regulation and consumer and labour protections. Data from RIA’s After Access surveys from 2018 and
2022 confirmed the replication of existing patterns of exploitation in these new work models .The
governance of these complex and adaptive systems is one of the most vexing policy problems facing
states in their efforts to improve social and economic inclusion. It is problematic not just from a
resource allocation point of view, but also because it is, unchecked, exacerbating existing socio
digital inequalities and even introducing new ones.

2.1. Digital inequality paradox

As affordable and meaningful access to digital services becomes critical to inclusive social and
economic engagement, and indeed to survival, as witnessed during the COVID-19 pandemic,
redressing the digital inequality paradox has become one of the most wicked policy problems of our
time. Efforts to ensure digital equality, not simply inclusion, have also become more complex than
they were a decade or two ago when policy concerns around the ‘digital divide’ reflected narrow
connectivity challenges resulting from lack of access to basic communication services. From a policy
and regulatory perspective, the paradox lies in the empirical observation that as more people are
connected, and as advanced technologies are layered over unevenly accessed and used foundational
technologies, digital inequality is increasing. This is not only the case between those online and
those offline (as is the case in a voice and basic text environment). It is also between those who have
the technical and financial resources to use the Internet optimally, and those who are ‘barely’ online.
The latter includes those who only have partial access to poor-quality or expensive data services that
do not permit them to be ‘always on’ or to use data-intensive services. The gap is widening between
those who passively consume a limited number of basic services, those able to put technology to full
and productive use, and the few able to innovate and contribute to the prosperity of nations. In
addition, the competencies to understand and use ICTs in the contemporary global political
economy are an essential element to consider, ranging from fluency in dominant languages through
to media and information literacies.

2.2. Intersectional inequality

Confronting the digital inequality paradox requires an in-depth understanding of the complexity of
these dynamic communications systems and a more nuanced and granular understanding of the
nature of inequality. Adopting an intersectional approach to understanding inequality can help to
overcome the homogenising language of marginality, exclusion and poverty and draw attention to
the relevance of social context, power relations, social inequality, relationality, social justice and
complexity. For analytical purposes, these concepts facilitate understanding of how digital inequality
is shaped by the intersection of multiple social identities, and the multifaceted points of policy
intervention required to redress or reduce it.
For example, social context is especially significant when importing policy reforms designed in the
Global North into the Global South, as their implementation may translate into different outcomes
from those experienced in the Global North. Likewise, examining social inequality at national and
global levels shows how power relations and social justice are linked not only to human rights
considerations but also to complex dynamics of global economic inequality. Indeed, while the digital
divide is the gap between those who do and do not have access to digital technologies, it more
importantly concerns the significant economic opportunity made possible through technology, as
well as capacities to affect governance around the control of such technology and wider digital
literacies.
Without economic or material justice, the rules may appear to be equally applied to everyone, yet
still produce unequal and unfair outcomes. Acknowledging these complexities provides the multiple
lenses necessary to conceptualise the dynamic and globalised processes of digitalisation and
datafication that we seek to measure and on which basis we seek to develop policy. An example is
that open access data is in theory open to all interested actors, but in practice can be put to most use
by powerful, private interests, unless proactive policy measures are applied.

2.3 Gender equity

One critical area of application is in gender equity. Based on the premise that the Internet can
contribute to the achievement of the Sustainable Development Goals (SDGs), calls for digital equality
have been foremost in organisational agendas both at the international and national levels over the
years. One of the precepts of the 2030 Agenda for Sustainable Development – “leave no one behind” – hinges on gender equality. SDG Goal 5b specifically identifies the enhanced use of enabling
technology, in particular ICTs, to promote the empowerment of women.

As Internet access figures increase, gender indicators will move towards the parity seen in mobile
voice services. But this is a long way off in most developing countries. It will also not happen
equitably, since those at the intersection of multiple inequalities are least likely to come online, even
as coverage increases, and prices decrease. Women in seemingly similar country contexts face very
different outcomes in terms of equal access to and use of technology. Even within countries, stark
differences are observed for women accessing the Internet, across factors such as geographic
distributions, education, age and business formalisation. The widespread abuse of women online
must also be taken into account, due to its likely influence in dissuading women (especially
journalists, activists and politicians) from full engagement online even when they have the technical
and economic capacities for access. The influence of these factors on access and use are themselves
not independent from each other either. There is a strong need for deeper analysis of nationally
representative individual-level data to better understand these intersectional inequalities, and the
linkages between technological developments and their potential to contribute to socio-economic
development. This will require systemically redressing underlying structural inequalities that are
mirrored in digital inequality. Equitable digital inclusion will require not only digital policy
interventions but integrated strategies to improve education and employment opportunities and
thereby incomes. This will ensure the right policy interventions which do not exacerbate current
inequalities. This is discussed further in the next section.

The implications of failing to address digital inequality in the information era are severe as the
Secretary General has pointed out. While global reform and donor agendas have shifted from
foundational digital inequality to issues of machine learning, AI for development and the ethics of
technical design, the ramifications and governance challenges of remedying inequality and exclusion
in the digital polity and economy are not unrelated. The exclusion of people from online financial
services, remote and platform work, and digital production makes them invisible in the data
extracted by global monopoly digital platforms for the purposes of creating lucrative digital
intelligence.
As a result, those at the intersection of multiple inequalities, of class, race, gender, citizenship and
ethnicity, are underrepresented and discriminated against in automated algorithmic decision
making. While proposed ethical and responsible design in data-driven technologies may go some
way to alleviating some of the obvious problems associated with bias and lack of transparency, they
are unable to deal with the structural inequalities reflected in the generation of data at scale and the
systemic perpetuation of historical injustices in their application to millions of micro decisions made
daily on the basis of giant social networking data sets.

2.4 Economic and social justice

While these structural inequalities will only be addressed through more fundamental economic and
social transformation, there are some systemic issues that can be redressed through policy
intervention. At the very least it will require the regulation of global digital public goods such as
spectrum, Internet, and data to ensure access to the means of communication and production and a
system of governance to mitigate against the associated risks. To promote more equitable and just
outcomes, economic regulation (as well as other regulatory arrangements) is necessary to enable
more even distribution of the opportunities arising from the data economy, not only the prevention
of harms to democracy and development. Awareness about the value of data for socio-economic
development and its ability to contribute to the realisation of the 2015 UN SDGs has become
increasingly prevalent.
With the global crisis precipitated by COVID-19, the growing dominance and linkages of data, big data
analytics, the Internet-of-Things (IoTs) and algorithms have placed data as a key resource in public
health management and economic reconstruction. This has amplified the need for data governance
and institutional arrangements to reduce the current unevenness of digitalisation and datafication
within and between countries. The emerging literature and practice of data governance have mostly
been approached from a negative regulatory perspective. That is to say it has sought to prevent
harms in relation to rights violations and mitigate associated risks – particularly privacy and security
but also freedom of expression. Positive discrimination to redress intersectional inequality in the
areas of access to affordable, adequate quality broadband, consumer protection, data protection,
public procurement and data access and sharing is required. Likewise, media and education literacy,
including digital political economy and digital rights, is also needed.
While various global and local epistemic communities are grappling with these issues, increasingly in
relation to AI becoming the next general-purpose technology, very little of this has focused on
economic governance. Yet there are many areas of data governance such as data availability,
accessibility, usability, integrity, as well as concerns about ownership, impacts on trade and
competition that require positive regulatory or governance intervention. Balancing current
commercial, supply-side valuation of data used in the allocation of resources and which has
produced these outcomes, with the demand-side valuation in the allocation of resources that
recognises their social value including as common goods is necessary to ensure more inclusive and
equitable policy outcomes.
Some of the reasons for this lack of attention to economic regulation (and to the neglected area of
demand-side value of pooled resources) relate the heterogeneity of data and the complexity of the
governance of different kinds of data. But as IT for Change has noted, “it suits those in whom much
of the value of data collected from across the world is concentrated for there to be no regulation …
Calls for the opening up of data markets or data flows without enabling the fair and equitable
participation of individuals, communities and countries disadvantage mostly countries in the Global
South and groups that are currently marginalised in the value chains and hierarchies from the global
digital economy …”
The pandemic and lockdowns have also amplified the scope for the private sector to fulfil traditional
public sector roles, or at least become significantly involved in their delivery without necessarily the
same accountability as the state. These partnerships create new power relations with which
substantive laws have been slow to engage. In addition, many developing countries are increasingly
reliant on digital giants in the Global North for the development of digital responses to a variety of
problems (e.g., contact-tracing apps’ reliance on a Google-Apple API), while a handful of digital giants
are playing an increasingly central role in various aspects of the digitalisation of all parts of life
leading to concerns of ‘digital colonialism’ and ‘surveillance capitalism’, along with a plethora of other risks that are unfolding in digital contexts (e.g., those related to abuse of digital dominance).
These present global regulatory challenges in which cross-border power evades national regulation.
This is especially problematic when such regulation is democratic, has adequate checks and balances, is participative and aligned to human rights, which are foregone by the cross-border operations of digital giants.

2.5 Public-interest research and public data

Despite such concerns, and the focus on preventing harms and mitigating risks associated with the
intensifying processes of datafication and digitalisation, research and systems of global governance
are still limited. The tools that do exist for evaluating the outcomes or implications of these
processes either tend to be applied to assess the collateral benefits of ICTs or to potentially negative
outcomes in primarily Global North contexts, rather than considering the likely very different impacts
in the very different contexts of developing countries. In the absence of public statistics and public
interest research, information is highly asymmetrical between citizens and communities and the
public and private sector, particularly big tech whose business is to extract data on a global scale to
create and sell market intelligence from which super profits are made. High-quality public data is
necessary to identify the precise points of policy intervention, the necessary regulation, and areas of
international cooperation required for governance. Public data is also required for planning and
implementation and to measure and assess policy and regulatory outcomes and the progress being
made towards a more equitable and sustainable planet. Without access to the swathe of data in the
private sector, regulation cannot be evidence-based, and policies for mitigation and empowerment
will be weakened.

3. The need for a governance framework in ensuring access to global digital public goods

Over the past three decades, the liberalisation and privatisation of traditionally publicly provisioned
communications services has driven innovation, creating an increasingly complex and adaptive
global digital ecosystem. However, regulation has always lagged innovation and market
developments, and is seldom flexible enough to enable market innovation while providing certainty
to long-term infrastructure investors, ensuring positive consumer welfare outcomes, and
safeguarding citizens from online harms to human rights. Further, the creation of regulated
competitive markets for the private provisioning of publicly provisioned goods such as the mobile
telecommunications, has not been extended to the Internet, or data-driven technologies and
platforms which have been largely unregulated. Although often presented as successful policy
outcomes of the traditional economic and competition regulators, mechanisms of universal services
funds to deal with market failures (or more accurately the market efficiency gap) have not been
successful. Large parts of the world lack the institutional endowments even to create and
effectively regulate private markets, much less the new forms of multistakeholder governance
required to manage these increasingly complex, adaptive systems. In addition, both driven by and in
response to intensifying globalisation, issues of cross-jurisdictional and global governance arose.
Even where these digital goods were developed in more mature economies, advanced technologies
have been treated as private goods, unregulated and highly excludable, despite having been
developed with public investments.
In response to the limitations of polarised state (public) versus market (private) ideological positions,
more pragmatic and appropriate forms of global governance have emerged. New forms of ‘public’,
best understood as practices or communities of practice and distinct from traditional state notions of
public, have emerged both in relation to the delivery and governance of public goods. These include
community stations, community networks, spectrum commons and free public Wi-FI and data
commons, and alternative forms of data stewardship such as trusts.
This has resulted in normative dilemmas that have both challenged and reinforced liberal
democratic norms and multilateral decision making, as global governance has been absent,
transformed or reconstituted.
The growing prevalence of collaborative forms of public goods governance (see below) does not
mean the end of the sovereign state; but it does change its nature and obligations in an increasingly
integrated world. In the state-market interplay, the state plays an important enabling role of
coordinating the private delivery of public goods and in ensuring the governance of global public
goods at the national level. Despite the extensive role of private and commercial delivery of
information infrastructure, with the increased positive externalities derived from digital and data
infrastructure (including data governance), the role of the state as coordinator and regulator in
ensuring its provision and management is still required to ensure widespread access by citizens, fair
distribution of social gains and the mitigation of risks.

3.1. Balancing demand side-valuation in resource allocation with
commercial, supply-side valuation of scarce resources

Regulatory and economic analyses of privately delivered public goods over the past three decades
have relied on the commercial, supply-side value of information, and financial and e-trade flows
with regards to the governance of global public goods. To develop alternative global governance
rationales, this submission makes the case for recognising the often ignored demand-side value of
public goods that are more likely to ensure digital inclusion in a developing country context. A
demand-side analysis focused on value-creation highlights that the outputs of digital infrastructure
industries are generally public and ‘non-market’ goods that create positive multipliers in both
economy and society.
This is why some classes of key resources (like broadband Internet, spectrum or data) need demand
side valuation as opposed to the usual supply-side valuations to properly recognise and account for
their public utility. Demand-side valuation, whether of broadband Internet, spectrum or of data,
enables public-interest governance of a resource as a non-rivalrous, low-excludability public good
that can be accessed for the purposes of public planning, entrepreneurship and democratic
accountability. Balancing the commercial valuation of resources necessary to ensure delivery of
certain public goods such as digital and data infrastructure with a demand-side approach, enables
the creation of a commons, allowing those who are unable to afford commercial services to access
spectrum through unlicensed spectrum or data through data lakes or alternative forms of data
stewardship. Doing so also enables policy perspectives to shift , for example, from narrow negative
regulation with rules and penalties focused mainly on compliance such as in spectrum management
or data protection to more positive regulation that redresses inequalities and enables participation,
to meet the objectives of digital equality and data justice.

3.2. Global governance of digital public goods

The rise of the Internet as a global digital public good underpinning global trade, financial and
information flows requires new forms of global cooperation. The shift in traditional power relations
between states, markets and citizens in global governance has blurred notions of ‘international’ and
‘national’ and of what constitutes public and private. After several decades of private interests
dominating evolving forms of data governance, the role of public regulation of the Internet and
specifically platforms has re-emerged as a priority. The current challenges to ensure the provision of
global digital public goods lie in the increasing complexity and adaptiveness of the global
communications system and the shifting global governance responses to these. These include
complementary and competing systems of governance ranging from nation-state-based multilateral
systems that have traditionally governed and coordinated global development, to new
multistakeholder formations accommodating state, private sector and civil society interests, as well
as to new forms of private authority, both commercial and non-commercial as found in ICANN.
However, Africa has been almost invisible in these developments.
The question arising is how Africa can better locate itself in global governance processes, not only to
ensure better outcomes in its own diverse interests but to participate more actively in setting the
global agenda. A key issue here is its capacity to do so. Understanding the Internet as an (impure)
global public good depends on Africa (and other regions) acquiring the relevant national and global
governance capacity to operationalise this understanding. This is because global consensus on the
good governance of the Internet (cybersecurity or data protection for example) only emerges, in
considerable measure, to the extent that countries can reproduce this consensus at the national (or
regional and sub-regional) level (e.g., creating the conditions for private delivery of public goods such
as the Internet, or complying with global agreements to enforce cybersecurity). Understanding the
Internet as a global digital public good can only be defended through the implementation at a
national level in all countries including developing countries of this understanding, often through an
imperfect global governance consensus.
Underpinning the policy and regulation of global digital public goods is that they are a common good
that has to be made available to all. While the concept of paying for national public goods such as
providing education or protecting clean air is widely understood, it is less clear who should be held
responsible for general-purpose global public goods, such as the Internet, that serve the common
interest. While investment in global public goods has traditionally taken the form of official
development assistance (ODA), new forms of international cooperation and institutions that will
support the development of global digital public goods and ensure greater digital inclusion are
necessary. This has produced highly uneven results. More effective and shared measures such as
better global resource mobilisation through digital taxes or other solidarity mechanisms that have
been discussed in this submission will be essential innovations to ensure the universal availability of
these common goods.
This submission applies this understanding of the need for effective governance of global digital
public goods to the role of developing countries in Internet, data and platform governance and the
development of public digital and data infrastructure. It draws on research that has used the concept
to track and explain the funding of overt technical assistance, and the tacit lobbying by various
interests (represented by multilateral agencies, global digital platforms, and industry associations) to
ensure the implementation of preferred global frameworks for cybersecurity, data protection and
data governance at the regional and national level. This submission also calls for greater
engagement and participation in global governance by African states, and in relation to their
potential for enforcing the legitimate taxation of revenues by global platforms that, without their
physical presence in countries, is unenforceable.

3.3. Recommendations

Digital and data public infrastructure – integrated broadband network infrastructure, the data and
services level and applications such as digital identification and payment systems – can only be
realised at the national level as a result of the global governance of global digital public goods. Even
if privately provisioned, the state needs, through public interest policy and regulation, to ensure
equitable access to public digital infrastructure so that what should be common infrastructures does
not serve a small elite segment of the population.

4.Critical themes

4.1. Public data for evidence-based policy to redress digital inequality

There is a severe lack of good quality and publicly available data that captures digital inequalities
faced by marginalised groups, particularly in relation to the needs of those who are offline and to the
situation in countries which have fallen furthest behind in digital access. Complex and nuanced
concepts such as gender tend to be reduced to narrow binary or categorical framing. Such
approaches fail to recognise heterogeneity within categories and the intersectional nature of
inequalities. Efforts to move beyond descriptive statistics allow for some demonstration of the
various factors determining the exclusion of people living at the intersections of multiple
inequalities. To appropriately inform and influence decision-making, it is necessary to produce
rigorous data differentiated along multiple dimensions to help isolate the exact points of policy
intervention required. To this end, through its nationally representative After Access surveys, RIA has
been producing high-quality data on digitalisation and datafication in Africa since 2004.

The After Access surveys show how households interact with ICTs across multiple countries in Africa,
providing valuable insights into digital inequalities, including with respect to gender. Data from the
2022 round shows contrasting gendered trends across different countries (Figure 1). In South Africa,
for example, Internet access has increased significantly since 2018 and the gender gap has been
reduced to a negligible level. On the other hand, in Nigeria, while the gender gap has declined since
2018, overall, Internet access has also declined. In Uganda, overall Internet access has increased but
mainly for men, with the gender gap widening since 2018.

Data Source: (RIA, 2012, 2018, 2022)

The survey highlights the risks of aggregating statistics on digitalisation both within African
countries, and for Africa as a whole. In all three countries there has been a significant increase in
Internet access at the household level since 2018, covering the COVID-19 period over which the share
of consumers reporting an increase in the frequency of Internet use ranged from 67% to 70%.
However, across all three countries, the increase happened almost entirely before the pandemic and
those who were not online before it occurred were unable to digitally substitute and will have been
further marginalised by the increasing prominence of the digital economy.

Similar considerations are pertinent for gender data. Gender is inseparable from race, class, culture
and religion, and cannot be understood in terms of discrete, quantifiable indicators to which it is
often reduced in descriptive statistics or digital indicators. The importance of considering
intersectionality is demonstrated in the Internet access rates for different groups of females (Figure 2
below). Whilst levels vary, there is a universal variation in access based on geographic location,
income and education, highlighting the heterogeneity of females. Moreover, those who face extreme
digital marginalisation are individuals at the intersections of these inequalities, for example, rural
females with low levels of education residing in low-income households.

Data Source: (RIA, 2022)

Whereas it is advocated within the UN Statistical system that ICT surveys be conducted on a regular basis, across most countries in the Global South there are almost no dedicated resources to do so and at best a few digital indicators included in national censuses or household surveys. In addition, much of the very limited quantitative research fails to assess the intersectional nature of marginalisation. The World Bank’s World Development Report (World Bank, 2021) notes that when data quality is poor, it lacks granularity, accuracy, and comparability. It also notes that the gaps in data on women and girls are particularly severe. Only 10 of the 54 gender-specific indicators (19%) in the SDGs are widely available and only 24% of the available gender-specific indicators are from 2010 or later.

The increased availability and usability of big data provides valuable opportunities for analysing
consumer behaviour, but from a policy perspective there needs to be more focus on analysing the
behaviour and needs of the marginalised and those whose rights to expression and access to
information are at risk. A major challenge in collecting data on marginalised groups and in
marginalised countries is that most individuals operate “invisibly” and, as they are not online or
minimally so, need to be identified so that the multiple factors preventing people coming online are
as well understood as the factors enabling people coming online.
Access levels are not the ‘be all and end all’, and it is important that digitalisation efforts ensure that
once individuals gain Internet access, they are also able to equally use the Internet meaningfully and
without fear. Rights must be respected online, as offline. In South Africa, the gender gap has been
eliminated at the national level. However, when viewing specific uses of the Internet through a
gender lens, it is clear that females are at a disadvantage when compared to men. In nine (75%) out
of twelve use categories, male use is higher than that of females (Figure 3).

Data Source: (RIA, 2022)

Whilst the After Access surveys provide valuable insights, they are limited to a small selection of
African countries and the contrasting situations and trends observed across these countries show that these cannot be considered accurate representations for the rest of the continent. Outside of these surveys, such data is in critically short supply.

In summary, despite an increase in the number of individuals who are able to access and use digital
technologies, there remain large gaps across different regions, across different countries within
regions, across different population groups within countries and even within population groups at
the intersections of other inequalities. Moreover, once analysis moves beyond access alone to assess
the ability to use digital technologies in a meaningful and safe way, even wider inequalities surface.
Whilst the COVID-19 pandemic accelerated digitalisation around the world, evidence suggests this
mainly came in the form of increases in intensity of use by those already online, with those offline
before the pandemic facing further marginalisation. It is also necessary to underline that even when
all conditions are considered for ensuring optimum access to the Internet, there need to be
conditions related to the supply side – such as local content online, independent journalism and a
free flow of information in line with the right to freedom of expression and access to information in
international human rights law.

Nationally representative individual-level data on ICT access and use is critical for identifying the precise points of policy intervention in evidence-based policy making. Without this, national indicators and assessments of users’ needs will always be biased towards more developed countries and towards those who have already been integrated into the digital economy.

4.1.1. Recommendations

Given the above, the following policy recommendations are made:
▪ Produce better data: The UN statistical system is currently unable to obtain the granular
data required for digital policy and planning from developing countries. An international
mechanism needs to be developed to ensure the collection of digital public statistics as a
public good, accessible and usable by all, and with all necessary ethical and rights-preserving
aspects of gathering large data sets. This should include efforts to support African use of such
data, including by supporting regional entities such as the African Union, UN Economic
Commission for Africa (UNECA) and the African Development Bank to coordinate statistical
collection. This will require enhancing the capacity of national statistical offices and political
commitment to do so. In addition, efforts are needed for enabling and tapping into academic
and civil society expertise on the continent to support the collection and analysis of
nationally representative demand-side statistics.
▪ Establish a digital solidarity fund: As a single key recommendation for the creation of
global digital data as a public good, it is proposed that there is a 1% contribution from the
DNS registration fees of all countries globally towards a digital solidarity fund. This could be
allocated based on applications by states to the fund to enable the gathering of digital data
that can be disaggregated, analysed and evaluated.
▪ End siloed ICT policies: Develop evidence-based transversal digital strategies focused on
human development, that foster more equitable social and economic inclusion, and which
include rights-preserving regulatory arrangements needed to redress intersectional
inequalities reflected in the uneven impacts of digital developments.

4.2. Connectivity, access, pricing and quality

The UN has declared access to the Internet as essential to the exercising of the right to freedom of
expression and opinion and other fundamental human rights, calling on states to ensure that
affordable broadband access is available to their citizens and access to the Internet is not unlawfully
restricted. Yet most Africans do not have access to the Internet, and for many of the few who do, their
use is highly constrained by the price of broadband data and the quality of services.

4.2.1. Connectivity

While telecommunications market liberalisation resulted in a communications revolution that
connected Africa to the world through multiple undersea cables and cross-border fibre networks
traversing most parts of the continent and brought basic telecommunications to the vast majority of
Africans, very few of this vast majority actually use the Internet. This is despite mobile broadband
coverage being above 75% in most countries and well above 90% in many. This problem with low
Internet usage on the continent is the result of both supply- and demand-side constraints.
Broadband coverage has been commercially driven with very few success stories of Universal Service
Funds (USFs) successfully dealing with the connectivity gap in areas where low rates of profit mean
that private telecommunications companies are not willing to invest in network extension in those
areas. Sometimes referred to as market failure, the gap has more accurately been described as a
market efficiency gap.

USFs are usually financed through mandatory contributions by telecom operators, with the intention
of the fund being dispersed back to the operators or other entities to support connecting the
unconnected and underserved. However, these have not delivered on the promise of better
connectivity nor have they been updated to include the new range of connectivity providers that
emerged over the last 10-15 years.

The equitable assignment of spectrum in a manner that best serves national strategic interests
remains a challenge. Spectrum auctions have proven difficult to execute well. High fees paid at
auction may be a windfall for the exchequer but have been shown to result in lower consumer
welfare through reduced network roll-out and higher consumer prices.

4.2.2. Pricing and affordability

Despite significant investments in network extension, these have often not been effectively regulated
to produce competitive markets and ensure positive consumer welfare outcomes. Even with the
price reductions documented in the RIA Africa Mobile Pricing (RAMP) Index we know from the After
Access that the main barrier to Internet take-up and use relates to the affordability of a smart device
and, once people are connected, the price of data. We also know from modelling data that what
determines access and intensity of use is education and, its corollary, income (employment). Women,
especially those at the intersection of multiple other inequalities, are overrepresented amongst
those without access.

We also know that it is almost impossible to deliver broadband services at less than 2% of the
monthly GNI per capita of least developed – and even several developing countries – as proposed by
the Broadband Commission. This is particularly not possible with the current high cost of GSM
technologies, inefficient business models, high transactional cost of spectrum assignment, and
regressive excise taxes on low-end smart devices, social media and mobile money.

RIAs After Access surveys show, in South Africa and Uganda, 70% and 61% cited data prices are the
main impediments to access respectively (RIA, 2022). The evidence of these high data prices is
highlighted in the RAMP Index database which tracks the cheapest data and voice/SMS basket prices
for each mobile network operator in Africa, on a quarterly basis. Figure 4 shows the average price for
a 1GB monthly basket charged in different African countries in Quarter 1, 2023. Research has shown
that the high data prices in African countries are due to a lack of competition in the mobile operator
markets.

4.2.3. Quality

On Internet quality, progress is evident. Coverage through 3G increased by 186% between 2010 and
2020, and 4G coverage has increased threefold in the last five years. However, the Speedtest by Ookla
shows that African countries still have on average low mobile and fixed Internet speeds. Even before
the pandemic, Internet speed was below the acceptable threshold of 10mbps, considered the lower
bound for a good quality broadband service.
Because of these serious access challenges to do with connectivity, pricing and quality, we cannot
continue to do the same things in relation to affordable universal access policy and regulation and hope for different outcomes. We propose a fundamentally different approach to redressing access, use and connectivity issues. Through more effectively regulated competitive markets that enable cost-based wholesale access to incumbent networks, and entry into the market of alternative and cheaper business models and technologies to meet the needs of all citizens, the infrastructure and services gaps can be met in a way that dominant operators are unlikely to. Considerably cheaper services could be provided through regional dynamic spectrum licensing particularly in underutilised rural areas and micro and community stations with access to shared or common infrastructure.
Regulation which served well in predictable, slow-moving markets is no longer able to keep up with
the pace of technological change and is often failing to enable smaller, nimbler organisations that
can address access challenges with new technologies and business models. Balancing the dominant
commercial valuation of spectrum with demand-side valuation that recognises it as public good, and
critical as a downstream input with both social and economic multipliers, is essential to achieve a more equitable global digital compact. This includes setting aside spectrum for common use and to enable free public access at schools and other public buildings.

4.2.4. Recommendations

▪ Build capacity among regulators: Regulatory frameworks are not keeping up with technological progress, the consequences of which are most obvious in spectrum management and administering of USFs. Therefore, it is essential for other stakeholders (such as civil society and academia) to be involved by:
-Providing training to national regulators, regional regulatory authorities, policy makers and public fund holders on the constantly evolving technology landscape.
-Working together with regulators and policy makers to study the impact of innovations on their own policy and regulatory frameworks, and update the frameworks accordingly.
-Reviewing strategic plans via public consultations, which can eventually lead to the creation of dedicated projects addressing a specific issue.
▪ Develop innovative regulatory approaches: Regulatory approaches to increasing connectivity are often designed to favour incumbents as the drivers of an expansion in access. Non-traditional connectivity providers and complementary access solutions, such as small operators and community networks, remain overlooked and often do not fit into existing licensing regimes. However, they are relevant actors particularly in providing last mile connectivity and need to be considered when the goals are connecting underserved areas and lowering the price of access.
▪ Public financing: Through evidence-based analysis assess the value and regulatory transactions costs of USF access funds in contributing to the achievement of affordable, meaningful access. If infrastructural deficits remain, create more agile and accountable USFs that reach the unconnected and underserved, while promoting competition and resiliency through a more diverse operator/provider base. The rules around governing and operating these public funds should evolve as technologies evolve. Alternatively, use the funds to support the demand-side barrier to adoption such as digital skills shortages. shift the levy for the provisioning of digital public goods to higher value contributions from operators and
platforms such as data. Mandate that public access to big data that could be used as administrative data and that currently sit behind a veil of unjustified confidentiality, be opened , even six months after market use for public policy and planning. Mandating contributions to data commons are all potentially of higher social value and easier to implement.
▪ Infrastructure sharing: Regulatory provisions for mandating infrastructure sharing can counter market concentration and minimise unnecessary redundancy and duplication of investments in supporting infrastructure (ducts, poles, towers, energy distribution grids).
This allows for reallocation of infrastructure costs to investments in service quality improvements, and reduces the market entry barrier for non-traditional connectivity
providers and complementary access solutions.
▪ Provide incentives for serving rural, remote, or underserved areas: Create regulatory enabled business models where small operators help incumbents achieve universal service obligations. Introduce simplified licence exemptions or licensing with low transaction costs for small operators and community networks.
▪ Promote spectrum sharing: Traditional operators have the rights to use portions of the spectrum nationwide, but only deploy their services in economically viable areas. New approaches in spectrum assignment should take into account recent innovations that enable dynamic spectrum access and spectrum sharing.
▪ Relinquish spectrum that is not in use: Develop administrative incentives to encourage licence holders to relinquish spectrum that is not in use (“use it or share it”). Give operators spectrum auction price or tax rebates on 5G spectrum licences for sharing their currently assigned spectrum.
▪ Allow wholesale access in the data market: To reduce anti-competitive practices in the data markets in Africa, allowing wholesale access, such as for the costs of roaming and facilities leasing facilities, would allow the late entrants room to compete.
▪ Prohibit of social media and mobile money excise taxes: These taxes are regressive as they impact on the welfare of people and the affordability of Internet access.
▪ Open data and transparency: For all the above mentioned recommendations to be efficiently implemented, there is a need for open telecommunication data (detailed information on existing infrastructure, equipment, spectrum assignment plans, fibre backbone etc.) as well as transparent information about the backhaul pricing.

4.3. E-Trade, labour, taxation and social protection

The UN Secretary General has called for a renewed social compact in the context of post pandemic
economic recovery. He has highlighted that human-centred recovery from the pandemic needs
employment and social protection policies to work in tandem, not only to improve people’s living
standards, but also to navigate the challenges of a rapidly changing modes of work (digitalisation)
and the transition towards the goal of net zero carbon emissions by 2050. With digital companies
making up the vast majority of global value add, they are also an obvious source for resource mobilisation for post pandemic economic and social reconstruction. The Global Digital Compact is a
complementary framework that can enable opportunities associated with these developments while
mitigating potential harms.

4.3.1. Trade

To unlock the value of digital trade for the benefit of all, there is a need to design inclusive global digital trade and data policies that acknowledges the critical role that African countries and labour play across the digital and data value chain, from the provision of scarce minerals to labour used for online content moderation. The African Continental Free Trade Area (AfCFTA) provides an enabling trading space for African businesses and industries to leverage each other’s expertise through regional value chains and to benefit from economies of scale and scope afforded by a single to integrated market. This enables Africa to be a far more competitive global player. Digitalisation is at the core of enabling this trade, and to facilitating its economic and welfare gains. African countries have come together to integrate their markets and move towards a digital single market under the AfCFTA. The primary strategy of the continent’s Digital Transformation Strategy for Africa is the Digital Single Market, which has a target of being achieved by 2030. The aim of the strategy is to develop common standards and a data protection framework that simultaneously aligns with domestic country contexts, regional requirements, and global standards.

This will require significant harmonisation and domestication of the African Union Data Policy
Framework which sets out a common vision, principles, strategic priorities and key recommendations to guide Member States in developing and integrating their national data systems and capabilities. This is so they can derive value from data that is being generated by citizens, government entities and industries. Currently, this data is appropriated by large multinational corporations, global monopoly platforms, and companies experimenting on or with data-driven technologies -sometimes as part of solutions for the public sector or humanitarian agencies, which is then commercialised for sale and profit extraction. Implementation and harmonisation of data policy (personal and non-personal), underpinned by a coherent governance approach which allows for the development of integrated data systems, is critical to optimise information flows and productivity gains from digitalisation and datafication on and for the continent. It is also important to support local companies, authorities and civil society initiatives so that they can provide solutions and expertise.
To unlock the value of digital trade for the benefit of all, there is a need to design inclusive global
digital trade and data policies that work for all Africans and in particular the most vulnerable. This
involves having more inclusive multistakeholder engagements that enable open and transparent
dialogue on how to regulate and govern digital trade so that everyone can realise the value of their
data. There is also a need for concerted efforts to ensure that digital trade under the AfCFTA is just
and equitable for all Africans, including informal traders and marginalised groups, who are often left
vulnerable and unprotected.

4.3.2. Labour

The digital economy, particularly location-based platforms, could potentially provide alternative
forms of work to alleviate mass unemployment in many countries across the continent. Digital
platforms can offer opportunities to people who face unequal access to income-generating
opportunities or endure systemic barriers to entry in traditional labour markets (Ahmed et al, 2021).
There is evidence that the presence of global digital platforms can benefit developing countries by
providing access to new technology, jobs, or skills. Conversely, these platforms also give rise to
factors that may amplify discrimination, resulting in job losses or declining incomes and inequitable
outcomes in traditional labour markets. It is important to highlight that precarious working
conditions characterise workers on these platforms (those who deliver the services) and that the
workers operate outside of official employee-employer relationships and are therefore unprotected
by labour law. ILO For instance argues that the type of labour relations mediated by digital platforms
spawn working arrangements that do not guarantee fundamental worker rights such as paid leave,
retirement, safety, and the right to be paid fairly for one’s contribution. Furthermore, workers whose
work is mediated through these platforms face overlapping challenges, such as an unfair bias against
women, labour supply exceeding demand, limited social protection, and income volatility. As we
have highlighted, this leads to a rise in corporate power and market concentration, and erosion of
labour bargaining power—which make it difficult for parties involved in these ecosystems to reap
equitable gains from the growth of digital platforms in the locations where they operate and earn
profits. Moreover, there is growing evidence that employment in the on-demand or gig economy can
replicate offline patterns of labour exploitation, particularly for lower-skilled individuals who manage
to find (inferior quality and low-paying) work online. Similarly, the engagement of low-paid workers
in the Global South for the purposes of content moderation and data-cleaning is an issue that call
out for global policy attention.

4.3.3. Taxes

Global digital platforms operating in developing countries may also evade taxes, as well as be non
compliant with national regulations. Historically, global processes of digitalisation and datafication
have been seen as a threat to the often-marginal tax bases that exist in developing countries. Tax
base erosion due to profit-shifting by digital platforms has been estimated to cost developing
countries over USD 500 billion annually. This points to the need to fix outdated tax systems, whilst
ensuring fair competition between multinationals and national businesses and that local start-ups
and smaller-scale initiatives are not obstructed.
The increase in the use of excise taxes on social media and mobile apps further highlights poor tax
administrative capability in developing countries (Matheson and Petit 2017). Some countries in Africa
for instance (e.g., Cameroon, Tanzania, Uganda and Zambia) have put in place laws that impose
regressive daily use taxes on Over The Top (OTT) services such as Facebook, Twitter, Instagram, and
LinkedIn, as well as on instant messaging apps like WhatsApp, Snapchat, and Skype. They often
implement these taxes without a comprehensive assessment of the long-term economic impact or an
empirical evidence base. Once introduced, they are seldom repealed. The procedure of calculating
these taxes is usually opaque, does not involve a public consultation process, and is justified by a
misguided understanding of the role of OTTs in the Internet value chain.
Not only do these impacts negatively on connectivity and affordability of mobile services, especially
for those with pre-existing inequalities who already face connectivity barriers, but also has a negative
impact on all segments of the Internet value chain, which hinders broader economic advantages and
digital development opportunities associated with ICTs. Our research shows that these excise duties
also have implications on the right to freedom of expression and access to information, which are
increasingly best exercised online. These taxes impact affordability and meaningful access to the
Internet, especially for the marginalised and poor. Ultimately poorly designed digital taxes have
actually lowered domestic tax revenue and reduced Internet use, as evidenced in particular in Uganda.
Regarding taxation, regulations should be crafted that address the abuses and monopoly or
oligopoly that is currently evidenced in labour markets. To mitigate the negative impact of social
media taxes, regulations that enable ICTs to be leveraged for sustainable economic growth and
social inclusion should be introduced. Should taxes be imposed, the process of their formulation
should be transparent, broad-based, and subject to a rigorous economic impact assessment.

4.3.4. Social protection

The digitalisation of the economy offers opportunities for the collection of much-needed tax by the
state as well as the ability to deal with challenges associated with enforcement through the visibility
of firms and workers. This creates more feasible tax bases for the social protection of currently
unprotected informal workers online and offline, and for more general social investment and
welfare. Social welfare interventions were relied upon extensively in Africa during the COVID-19
pandemic, as governments turned to digital cash transfers as a preferred tool to protect the poor and
vulnerable from the socioeconomic impacts of the pandemic. In South Africa for instance, 18% of
households received the government’s COVID-19 relief. Of these grant recipients, 70% were able to
receive grants through digital channels. The percentage of the population that received grants
digitally was much lower in Nigeria at 3%, where only 32% of households were eligible to receive
grants through digital channels.
As a result of lockdown policies that were instituted across the continent in relation to social
protection, there is the potential to make regulations to support digitally enabled innovations for the
intake and registration of beneficiaries, assessment of needs and conditions, and provision of benefits. Digital transformation on the continent therefore offers an unprecedented opportunity to address eligibility and barriers to formalisation faced by informal individuals and firms if the benefits of doing so are evident eg. business disaster relief, online training, online financing, micro business procurement by governments

4.3.5. Recommendations

Regulation should target the negative labour market distortions that digital platforms can generate.
However, addressing the complex adaptive global nature of platform work requires an ecosystem
approach. This means that all stakeholders who are part of the location-based platform ecosystem
(platforms, regulators, the government, platform workers and trade unions, and consumers) should
be consulted to craft policies that create the fair working conditions needed to promote both gender
equity and economic justice for platform workers. They should gear this regulation toward
rebalancing bargaining power, extending certain labour rights and protections, and reshaping social
protection provisions.
In light of this, the following recommendations are made:
▪ Pay attention to the knowledge lag: The evolving nature of the ICT landscape needs to be better understood. A number of traditional business and network models, regulatory frameworks and infrastructure investments in the African ICT landscape are critically out of sync with the fast pace of new ICT developments. Policy makers should keep abreast with these dynamics and develop localised solutions that ensure that fiscal measures do not contradict digital development strategies with long-term economic and social inclusion implications.
▪ Address structural constraints: Address the key structural constraints that inhibit online labour rights, gender equality and inclusive social protection. Concerted efforts are needed to ensure the interventions that are designed to strengthen economic justice in the digital economy simultaneously facilitate a policy and regulatory environment which is fit for purpose and contextually relevant to address existing labour market inefficiencies that are exacerbated by location-based platformisation.
▪ Recognise jurisdictional challenges: Acknowledge the jurisdictional challenges of enforcement when digital platforms do not have any physical presence in the country, governments should cooperate to ensure that (multinational) location-based platforms provide a minimum set of protections to their workers, such as facilitating working conditions that align with existing relevant national labour regulations.
▪ Create trusted working conditions: Compel location-based digital platforms through global governance to provide more transparent and innovative communication strategies and algorithms to facilitate robust and trusted working conditions.
▪ Align fiscal and tax regimes: Align national and regional fiscal and taxation regimes with efforts to mitigate against Base Erosion and Profit shifting (BEPs) to effectively tax digital platforms.
▪ Enable better distribution of tax between countries: Foster cooperation for the distribution of taxes between countries. Leaving it to states to facilitate these agreements can contribute to further imbalances across the continent, as disparities will emerge from current patterns of collaborations where weaker countries are left out.
▪ Support multistakeholder engagement: Create ongoing, inclusive multistakeholder engagements that allow an open and transparent dialogue on how to improve digital trade so that everyone can realise the value of their data towards a public good, with due protections and benefits.

4.4. Data governance and realising data justice

Data has, over a relatively short period, emerged as a defining force of social organisation and of global capitalism which can be harnessed to influence and govern almost every aspect of our day-to day lives. While progressive conceptions of data governance are emerging in Europe, legal and institutional responses have for the most part drawn on a narrow conception of data rights at the individual level, focusing chiefly on personal data protection and privacy. Even so, personal data remains almost completely unregulated in many parts of the world, as does non-personal data. Yet, when accumulated at scale, data confers almost unimaginable power to extract value and govern peoples’ lives.

Without regulation and other interventions, data has become a site of injustice— exacerbating
inequality and marginalisation. Regulation at national and supra-national levels attempts to tackle
some of these harms. However, in addition to the narrow focus on personal data and individual
rights, regulation has been formulated in a broad and general way, overlooking contextual specificities and not providing sufficiently focused implementation, accountability and enforcement frameworks.

These measures largely fail to respond to the potential harms of datafication which take place at the
collective level, as well as the role of data as a resource and a means of production which confers
enormous economic power when accumulated in large volumes. To address this gap, policy
approaches are needed to govern data in the collective interest or for common good. These must be
guided by overarching principles of equity and justice; however, as Parminder Singh and Anita
Gurumurthy from IT for Change contend, these principles must also be responsive to local and
contextual differences in how datafication is experienced.

To advance this goal, RIA participated in developing the African Union Data Policy Framework to
provide an Africa-centric data justice-focused governance structure for the burgeoning data
environment across the continent. Building on the theoretical framework of both economic and data
justice, we argue for a strong incorporation of these concepts into the GDC. Considering the limitations of comparable data governance frameworks, the example of the African Union developing a data policy framework that incorporates data justice, despite inadequate public participation in the African Commission and Member State adoption processes, is an example of how these important principles can be incorporated into a high-level compact.

4.4.1. Data governance

At the heart of data governance is instilling trust in the use of data as a pre-condition for fully realising the gains of digital transformation. In response, many have called for “data governance” approaches to enable institutional and regulatory structures to govern data in line with normative frames that assume:

• democratic and rights frameworks,
• institutional endowments to govern effectively, and
• levels of human development that allow citizens to exercise their rights and freedoms.
  
  However, these conditions do not pertain in many countries toward which the SDGs are specifically
  directed and there is now considerable evidence that regulatory governance approaches premised
  on these assumptions that work in the Global North do not always apply in the Global South. As
  Linnet Taylor has argued, not only are such approaches sometimes insufficient to create the conditions for truly beneficial data and AI in the Global South, but they also fail to account or prepare for the reality of contextual developmental demands – along with the risks that are introduced by the increasing production of digital traces, shadows and selves.
  
  Any examination of the data realities into which AI is being introduced requires a wider political
  economy assessment with consideration of the global and domestic power relations, interests and
  imbalances that determine technological developments. Technological advancements, especially
  when coupled with large-scale crises like COVID-19, have highlighted some of the practical
  considerations for data governance in the public interest and some of the normative tensions in the
  treatment of private, community, and collective information that have clear implications for AI.
  
  Currently, control of data is concentrated amongst a small number of platforms primarily in the US
  and China. In Africa, private actors such as telecoms companies along with banks, direct messaging
  services, satellite operators, the retail sector, and mines, etc. have become major holders of data.
  States have been lagging in their role to gather, process, utilise and avail data to external
  stakeholders, although the number of African countries with Right to Information laws has increased
  from just five in 2009 to 30 in 2023 according to UNESCO. This underscores the need for economic
  regulation to enable greater access to and control over the resources of data and AI, and the
  capabilities and infrastructure required to utilise them, for more even redistribution of opportunities
  and benefits within and between countries.
  
  Furthermore, data collection and data-driven systems are already exacerbating social, and economic
  inequalities, as they amplify bias, marginalisation and discrimination against vulnerable or
  disadvantaged groups (such as women). This demonstrates the need for rights such as equal
  representation in data, non-discrimination, and the protection of group identities and indigenous
  knowledge systems. Some practices, such as the use of biometric data for predicting proclivity for
  crime, are inherently unjust, cannot be resolved through improving technology and must never be
  permitted.
  
  4.4.2. Data justice
  
  The concept of data justice has emerged in response to these critical concerns. It has been advanced
  by a global community of researchers and activists such as Taylor as a way to talk about and challenge the complex, overlapping harms of datafication at multiple scales – personal, local, and
  global. There is also research beginning to link data justice to economic fairness and rights and to
  create comprehensive frameworks for what economic justice and a fair distribution of benefits might
  look like in a datafied society. Data justice requires that data are not used to sustain or compound
  existing inequalities or undermine autonomy, that those from whom data originates share in the
  economic and social value realised from data, and that there is an equitable distribution of
  opportunities arising from data. The very conception of what constitutes data, as a meaningful set of
  signals that can be processed, needs attention. Definitions of data can impact adversely on issues
  like the need for limits on bulk surveillance, facial recognition, gait recognition, and gender.
  
  Expansive approaches are needed to govern data in the collective interest, or for the common good,
  while still protecting individual privacy. Recognising the validity of communal ownership of data and
  the impact of data use on communities, data justice requires protection of collective rights in data,
  equitable access to economic gain from collective data, and community protection from harms, such
  as bias and discrimination, caused by data. It also demands attention to how digital infrastructure is
  owned and controlled by the northern hemisphere which, in turn, determines where data, power and
  capital will be concentrated. The advancement of AI and data have shown the economic power
  derived by online platforms which extract data from users, converting such data into “digital
  intelligence” that enables them to develop new products and services and increase efficiency in their
  profit-output operation, to the detriment of users who gain little or no economic value despite being
  the producers of this data. Also significant is the training of AI on data that is copyright protected,
  such as data behind media paywalls, and where data is scraped from public sources without
  permission. The risk for Africa is that the costly production of data may benefit AI applications which
  contribute nothing to the generation of the data resources on which they rely. The ambiguous issue
  of AI operators taking advantage of data stored in cloud services, or gathered under auspices such as analytic or cybersecurity services, needs attention.
  
  The Global Partnership on Artificial Intelligence (GPAI) data and economic justice primers raise
  concerns about the increasing power of a few digital monopolies over digital infrastructure. The
  African Union Data Policy Framework also notes the Global North/South imbalance in the
  distribution of platform owners and their tax jurisdictions. This imbalance is enabled by the lack of a
  framework on moral and legal principles of economic justice. Consequently, in anticipation of the
  future development of the digital commons as a global public good, and as elaborated in GPAI data
  and economic justice primers on which we collaborated, as well as in line with the collaboration
  with IT For Change on a feminist digital economy, we support four rights-based principles for data
  related economic justice. These are:
  ▪ the right to benefit from one’s data, and avoid economic harm;
  ▪ the right to access one’s data, including for third parties of choice;
  ▪ the right to appropriate representation in data, including the right to invisibility and remaining absent, and;
  ▪ the right to participate in the governance of data, and in the relevant economic systems based on data.
  In addition, economic data rights should be recognised for data generators and data workers, who
  can use their collective rights to port aggregated data to a data commons or alternative data
  infrastructure. This could foster competitive data-based services that could serve collectives much
  better than data monopolies do. Rights alone do not ensure economic justice, however, they should
  be instilled in relevant laws, regulations, policies, and implementation mechanisms, to enable
  equitable and just economic outcomes.
  
  4.4.3. Recommendations
  
  Key policy recommendations coalesce around localising and contextualising data rights and
  protections; ensuring affected communities can participate meaningfully in data governance; and
  guarding against the extraction and concentration of data as a resource and means of production.
  The following high-level policy recommendations are applicable at different levels of government, as
  well as in non-government and private sector settings. Within specific country and regional contexts
  and through the participatory processes intrinsic to democratic policy formulation, more detailed
  guidelines for achieving data justice can be developed.
  ▪ Ensure data governance and regulation uphold rights, including economic rights and workers rights such as the right to port their work-related data to other platforms, and appropriate frameworks for redress that bind transnational corporations.
  ▪ Contextualise and localise data rules through the democratic participation of affected communities and enabling alternative forms of data stewardship.
  ▪ Enable beneficial data flow by preventing anti-competitive data practices, ensuring equitable
  access to resources, and enabling interoperability.
  ▪ Reconcile the claims of data sovereignty and global governance through international solidarity committed to data justice including redress of inequality and marginalisation.
  
  
  4.5. Disinformation and information disorders
  
  4.5.1. Information disorders in Africa
  
  Reflecting some of the challenges at the global level, laws tackling information disorders in the continent tend to be ambiguous, with the potential of politicising the concept and criminalising
  speech in efforts to control narratives. As with other regions in the globe, elements of information
  disorders, such as disinformation and hate speech, manifest heavily during elections and conflicts.
  This has had serious implications in the exercise of democratic rights, freedom of expression and
  rights to political representation by structurally marginalised groups. Ideally, platform companies
  should collaborate with local authorities and stakeholders to deal with these challenges. The
  safeguarding of citizens’ fundamental human rights, as well their civic rights is not something that
  can be left to the discretion of providers of public services with the potential to cause significant
  harm. In the same way that international air transportation networks, whether private or public, are
  regulated to guarantee passenger safety, global communications networks need to conform to
  universal democratic rights of privacy, access to information and freedom of expression.
  
  While states in Africa have deficient regulations to properly address and account for the information
  disorders produced largely online, these systems of criminalisation are likely to prove even less
  effective through the likelihood of an increase in AI-generated false content, with its potential
  disruption to economies and political processes. Simultaneously, efforts by platforms to deal with
  the magnitude of the problem through automation of the identification and removal of fake or
  harmful information through the use of opaque algorithms, the training of which happens through
  the manual identification of harmful content by people employed under questionable labour
  standards by platforms, are not only inadequate but potentially harmful themselves. Because the
  staff of these global social media platform companies are spread too thin across the continent to
  offer contextualised solutions to information disorders, policies tackling information disorders, and
  consequently content moderation through algorithms are done from the perspectives of the Global
  North. Contributing to this challenge is the dearth of knowledge from the Global South. Studies on
  information disorders remain dominated by theoretical paradigms, examples, and case studies
  drawn from relatively recent experiences in Global North contexts.
  
  4.5.2. Disinformation and content regulation
  
  The proliferation of online hate speech and disinformation has serious implications for human rights,
  trust and safety, not least in African countries. “Information operations” around the continent aimed
  at covertly swaying elections have been widely exposed, while violence has been incited against
  ethnic groups and migrants in a number of conflict cases. A RIA study has detailed range of other
  problems with the integrity of Africa’s information environments.
  
  In another RIA report, commissioned by UNESCO, we have addressed platform regulation challenges
  that have significant negative bearing on Africa. These highlight the discriminatory application of
  platform regulation policies by major platform operators:
  ▪ platforms are slow to respond to emergencies that result in serious human rights violations, and
  ▪ they have failed to invest in African languages for moderation and, consequently, to provide
  linguistic contexts for content removal.
  
  Findings by RIA are that the key factors accounting for the proliferation of disinformation and hate
  speech online are:
  ▪ ‘attention economics’;
  ▪ automated advertising systems;
  ▪ external manipulators;
  ▪ company spending priorities;
  ▪ stakeholder knowledge deficits, and
  ● flaws in platforms’ policies and in their implementation.
  The research also finds that how platforms understand and identify harms is insufficiently mapped to
  human rights standards, and there is a gap in how generic policy elements should deal with local
  cases, different rights, and business models when there are tensions.
  In the absence of companies having clear elaboration of how their policies operate at global and
  local levels, it is to be expected that there will be many international inconsistencies in application.
  The GDC needs to account for the range of interactions and applicability of regulations both on
  global and local levels. Platform regulation should emphasise the value of clarity and transparency
  regarding application on both global and local levels – of platforms and of their different internal
  policies.
  Also important to consider is how platforms’ own policies on content regulation are generally
  overshadowed by content curation algorithms that support the core business aims to collect data
  and sell advertising.
  Further, platforms’ policies are not always transparent and that they do not provide adequate risk
  assessment. Policies that are shaped in this way, and in secrecy, work against the human right to
  equality.
  Challenges around transparency are presented in two ways: opacity in terms of lack of consistency in
  how policies are applied – for example, Google transparency reports on political advertising during the elections still misses data from the African continent – and opacity in how AI-based content
  moderation works.
  Enforcement by platforms of their own terms of service has grave shortfalls, while attempts to
  improve outcomes by automating moderation have their limits. These inequalities in policy and
  practice abound in relation to different categories of people, countries and languages, while
  technology advances are raising even more challenges.
  In this light, this submission suggests that the GDC advises that regulatory arrangements should
  include the interplay between platforms’ policy rules, practices, business models and technology. At
  the same time, the failures of ‘solo-governing’ by platforms in content curation and moderation,
  should not lead to ‘solo-state regulation’ as the governmental response. There are many examples in
  Africa of inappropriate regulatory responses to developments on platforms, in particular Internet
  shut-downs and criminalisation of content under vague terms such as “false news” or “fake news”.
  What speech and narratives get to be or not be accepted online should not be left to social media
  platforms, or one stakeholder to decide and regulate.
  Accordingly, the GDC should encourage hybrid regulatory arrangements that include multi
  stakeholder participation in the making of rules for and by platforms. The specifics may vary
  according to the issues at hand (e.g., elections, child protection, public health), but there should be
  cross-cutting requirements for transparency reporting, access to platform data, and conducting
  human rights impact assessments.
  In the light of our research, this submission suggests that the GDC advises that any evolving
  regulatory arrangements include the interplay between and within platforms’ internal rules,
  practices, business models and technology.
  
  4.5.3. Recommendations
  
  Tackling disinformation and hate speech online can best be done on a modular basis as befits
  differing issues, regulatory arrangements and capacities, and by ensuring that the array of statutory
  regulators is structurally independent of political interventions. To address disinformation and hate
  speech online:
  ▪ The GDC should highlight human rights and sustainable development as agreed international
  standards that are foundational for all regulatory arrangements everywhere.
  ▪ Statutory authorities should not seek to take over the formulation of platforms’ content
  policies nor the moderation work by the companies. Instead, they should require companies
  to meet their own consumer terms of service to the full, as well as to follow broader objectives, policy standards and process benchmarks.
  ▪ These principles can apply to solo-, self- and co-regulatory mechanisms, as well as be
  institutionalised in multistakeholder roles at all levels of rule-making, enforcement, monitoring, oversight and review.
  ▪ It can be strongly encouraged that platforms elaborate how they balance global and local
  dimensions of their terms of use, and provide more equitable (and auditable) resourcing for
  monitoring and moderation of content in the Global South..
  ▪ Tackling disinformation and hate speech online can best be done by combining various
  regulatory arrangements into a hybrid overall system including the legal delegation of roles
  to platforms, effective codes of conduct, ensuring statutory regulators are structurally
  independent from political interventions, and institutionalising multistakeholder involvement throughout.
  ▪ Media, NGOs, tech employee bodies, whistle-blowers and researchers should be supported
  as positive elements in the wider governance ecosystem in which regulatory arrangements take place.
  ▪ Responses to hate speech and disinformation should build stakeholder knowledge through media and information literacy, and support for de-centralised platforms.

4.6. Harnessing AI for Africa

The development and use of AI in Africa should support, rather than hinder, socio-economic justice
and poverty alleviation, and not impact negatively on the enjoyment and realisation of all human
rights. Yet, the foreign global monopolies that dominate the AI market in the region are based on
systems developed outside Africa and primarily trained with foreign data, thus limiting their capacity
to address the region’s priorities. Even worse, these foreign-trained models may have serious implications for human rights through models that are discriminatory or simply inappropriate for local contexts.

As AI rapidly proliferates on the African continent, it is essential to assess which regulation exists and
identify what new regulatory regimes are required to address the risks and guide the application of AI
for public good. RIA’s analysis of the dynamics of AI use and regulatory environment across the education, health and well-being, labour, climate change and environmental sustainability sectors in
Southern Africa indicates that significant efforts are required to ensure AI project leaders understand
and take into account the potential and ethical and human rights-related implications of their
products and projects. Further, the majority of AI use in the region is by privately-owned companies,
suggesting the need for efforts to support public sector capabilities. Uses also fall mostly into the
labour policy arena, indicating a need for the prioritisation of policies and programmes that address
future-of-work scenarios such as reskilling and labour transitions.

Though some African countries, such as Mauritius, Egypt, Rwanda, Ghana, South Africa and Tunisia,
have developed or are considering developing AI strategies, a majority of them do not have national
AI policies or strategies in place. Cooperation between countries and regional bodies, and between
countries and tech companies should be encouraged to increase the capacities of local authorities to
effect these policies for the protection of fundamental rights and freedoms, as well as for the promotion of AI deployment in the region. Rapid development in AI capabilities without an adequate legal and regulatory framework threatens to upend the livelihoods and rights.

4.6.1. Data governance towards innovation of AI in Africa

One of the challenges limiting the democratisation of AI and its deployment in the region by local
providers for local solutions, is the domination of both global and national tech monopolies who
have amassed huge control and power through massive data collection and infrastructure ownership. A lot of local solutions are bought by these companies because of their financial might.

Another challenge is that African governments lack open data initiatives that can be leveraged for AI
development. In countries where these initiatives have been set up, they have been unsustainable as the systems are not fed with data. Globally, the continent has the least statistical capacity. The lack of data deeply entrenches inequality and harms caused by AI due to lack of (local) data on marginalised and underrepresented groups that can be used to avert these harms. Contributing to this challenge is the low affordability of ownership and use of digital devices and networks. As Africans miss out on the opportunity to use these technologies, they also miss out on the opportunities to create a digital footprint, which is needed as raw material for AI.

4.6.2. AI and biometric identification

The application of AI in the area of biometric identification can make citizens more vulnerable in the
online space. AI-driven digital ID systems are rapidly being adopted by countries in Africa. These
systems entail the use of biometric data such as fingerprints, iris scans, and facial images for authentication in access to services. The rationale behind the adoption of these systems is to increase efficiency in provision of government services and to protect citizens against identity theft. However, these systems are susceptible to cybersecurity attacks and third-party sharing of data that leaves citizens vulnerable to data breaches. We are currently working on this project on Digital IDs in Africa.

4.6.3. AI and labour in Africa

It is important that the whole AI lifecycle respects human rights, human decency and autonomy.
Africa contributes significantly to the global AI supply chain: including as suppliers of raw materials for technology, producers of data, human content moderation supplementing algorithm-based moderation, cleaners of data sets, and agents feeding machines with content for machine learning. However, the current practices of combating information disorders through algorithmic as well as human content moderation are characterised by global hierarchies of exploitation. The inequalities in the distribution of labour exist both in the employment of human intervention to moderate content on social media platforms and the back-door labour of training AI to identify elements of information disorders. Both need to be seen as work that is distributed to lower income countries seemingly requiring little or no skills. Unfortunately, this makes the workers easily replaceable and leaves them with limited bargaining power and social protection. Therefore, platform regulations need to account for global hierarchies where workers situated in the Global South are doing work for users based in the Global North.

4.6.4. Recommendations

Given the above, the following policy recommendations can be made to create an enabling environment for AI in Africa:

▪ Infrastructure development: Governments should prioritise developing policies that ensure safe, secure, and inclusive infrastructure for the development of AI systems. The policies should also include means of fostering open data (which is very crucial for the development of AI systems), connectivity and Internet access, and good governance for the development and sustenance of AI systems.

▪ Building local capacity and skills: Skills development is at the heart of advancement and responsible use of AI in Africa. Policies should be formulated to promote understanding around AI at all levels, and promote uptake of STEM by women and underserved groups. Policy makers must also be supported to understand the risks of AI, and women should be central in AI-related decision-making positions.

▪ Taking a gender and intersectional sensitive approach to address current inequalities:
Expand gender-sensitive information and data collection for the broader ICT sector to inform policies that can advance and ensure the active leadership and involvement of women in digital and AI services. Standards around data quality and representation are essential to eradicating discrimination against women, and other underserved groups, that occur in AI systems trained on bias and unrepresentative, or under-representative, datasets.

▪ Advancing African-centred value systems in AI ethics: It is essential to address ethical concerns surrounding AI deployment. Governments, industry players, and other stakeholders should play a key role in creating regulatory and policy frameworks centred in African values and principles.

▪ International development assistance: In advancing development of responsible AI solutions in Africa, actors including donors, intergovernmental organisations and other funders should focus on supporting infrastructure development to ensure inclusive AI for its long-term development. In doing this, focus should also be placed on “ensuring African states retain their sovereignty in developing AI governance solutions” that are rooted in national values.

▪ Collaboration and knowledge sharing: There is a need for collaboration and knowledge sharing among countries and regions to promote the global deployment of AI for public value creation. African countries can benefit from partnerships with more advanced economies to enable them to leapfrog development stages and accelerate AI adoption.

▪ Accountability for risk: A governance framework for AI in which the handful of corporations
driving rapid AI development are required to account to public authorities for the current and
future risks presented by the technologies that they are developing and are required both to
manage risk and to maintain international trust.