Border politics: closed to humans but open for digital health data

AI is perceived as having the potential to transform healthcare access globally, for example, by using machine learning to make predictions and personalise treatment strategies and using data models and algorithms to analyse medical scans and pathology images.32 As such, AI relies on massive, unrestricted flows of data and complex technological infrastructures to collect, store and analyse it. Corporations that train AI models for health are increasingly looking for electronic health records from all over the world, especially in the global south, where this data has been difficult to collate.33 For many in the global south, this involves the extraction of digital health data by corporations primarily located in the global north. The extraction of health data for AI is also facilitated by powerful international organisations, such as non-governmental organisations like the WHO and the United States Agency for International Development (USAID). These organisations are investing in massive electronic health record systems that will ultimately be used to develop global health data models.34 Underlying these systems is an ideal of data flowing freely across national borders. By contrast, the free movement of people from the global south to the global north is becoming even more difficult through stricter border controls.35–37

These border controls have a long history in global health, with countries in the global south historically being linked to vectors of disease. Using international law, countries from the global north attempted to stem the flow of citizens from the global south through international sanitary conferences from 1851 to 1938. When citizens of the global south were allowed passage, this was often contingent on strict quarantine rules.38 During previous pandemics such as Ebola, we saw the dual standards with which these border policies were applied, with citizens from the global south locked out while those in the global north were allowed free entry.36 The Omicron variant during the last COVID-19 pandemic brought this discrimination into sharp focus when many countries in the global north closed their borders to South Africa and other African countries after South Africa proactively sequenced the variant.37 Additionally, the continuing lack of human mobility continues today, with many global health scholars from the global south being refused entry to the global north for conferences and other international health collaborations.35 39

Infrastructural colonialism through software, hardware and the cloud

Digital technology infrastructures, which include software, hardware and the cloud, are mainly dominated by a few conglomerates that are in the global north. Big tech corporations exert tech hegemony through racially extractive models of appropriation and exploitation of human life through data for economic gain. Digital health applications rely on an entire ecosystem that is protected through intellectual property rules.40 While there is a rhetoric of free software for technical developers such as Java, C++ and Python, many developers soon discover that they often need more sophisticated paid-for software in order to create sustainable applications. This is further compounded by the fact that digital application marketplaces such as the Apple App Store or the Google Play Store act as gatekeepers of code and often place software-related demands on developers who want to publish applications on their platforms.41 The process of application development also relies on increasingly powerful and expensive proprietary hardware in the form of computers, storage and graphics, which may be protected within the intellectual property (IP) system using copyright, trademarks and designs. Within AI, while free datasets have been met with great enthusiasm, which conveniently sidesteps that in the long term, many developers will still need to pay substantial amounts to access relevant AI models, which are all located in the global north.42 43 This is despite the fact that the AI models are trained on global data sets, many of which were acquired freely. Additionally, the use of secure servers for storage is restricted due to IP rights and proprietary information embedded in them, making it difficult for emerging global south companies to leverage these innovations for digital health application development. Thus, the promise of rhetoric that everyone can gain access to these digital tools and democratise digital health application development remains a fallacy, especially in global south contexts.

The privatisation of software and hardware is sustained through a global intellectual property system that allows companies in the global north exclusive rights of use over technologies. Under the 1995 Agreement on Trade-Related Aspects of Intellectual Property Rights, commonly known as TRIPS, inventors have exclusive rights to profit from their innovations for a period of at least 20 years. Digital health technologies usually include software and hardware. Patents can be used to protect the hardware in health-related systems such as sensors but some inventions relating to methods and protocols, such as software, may also be patentable if they are perceived as 'new and involve an inventive step'. This system of IP rights has created colonial legacies in access to healthcare.44 These colonial legacies have perhaps been best illustrated recently through the way IP control resulted in inequitable access to vaccines in the wake of the COVID-19 pandemic, with countries in the global north hoarding more vaccines than they needed while those in the global south struggled to secure them.38 45

Additionally, many countries in the global south depend on the global north for cloud services. With data centres located in the global north, the influence of big tech has filtered into global south government operations, some of which depend on external digital infrastructures such as cloud storage.46 Cloud computing can be described as ‘a model for enabling convenient, ubiquitous on-demand network access to a shared pool of computing resources.’47 The cloud has revolutionised storage and backup facilities and can be used to enable the sharing and accessibility of health data globally. Many of the earlier concerns about ‘the cloud’ were with security and privacy. However, the rise of end-to-end encryption has resolved some of the concerns, although global north countries such as the USA with its Cloud Act of 2018 can still request all USA cloud infrastructure providers to hand over data to law enforcement even if the location of data centres is abroad. The top cloud computing companies are still based in the global north, with the majority in the USA and Europe.48 Many countries are so dependent on this cloud infrastructure that even simple transfers between entities in the same country will often be reliant on a cloud that is often based in the global north.49 This means that even local companies in the global south using health apps for local purposes will be reliant on remote storage and backup facilities.

Philanthropy, aid, marketisation and the distortion of public health agendas

With advancements in digital health innovation, thanks to big tech, funding instruments, capitalistic logic and market-driven dynamics are reshaping public health agendas.50 Venture capitalists fund start-up innovation with the sole purpose of making the highest returns possible and digital health has emerged as a potential field for commercialisation.51 This has meant that financial returns are prioritised over health gains, which can distort public health objectives and outcomes. Additionally, the increased focus on digital capital at a cost in the global south entrenches existing inequalities, stigmatisations and discriminations along intersectional lines and reinforces colonial relations of dependence and dominance.52

Private funding for public services has always revealed how certain agendas get prioritised over others.53 This approach to digital health funding disproportionately affects countries in the global south that might not have adequate resources to invest in digital health infrastructure for their national health agendas. Some of the global south’s digital health programmes are donor funded by international non-governmental organisations located in the global north.15 These organisations influence health outcomes and also participate in or are complicit in sustaining the cross-border migration of data. For example, most of the donor-funded projects, such as the MEASURE Evaluation’s Data for Impact project (D4I)—a USAID-funded project that works primarily in Sub-Saharan Africa to increase capacity to use data for improving health programmes54—do not explicitly state how they manage issues of data governance such as stewardship of data, location and mode of data storage and what sufficient safeguards exist for ensuring appropriate use of data.55 MEASURE Evaluation D4I is a USAID-funded project based at the Carolina Population Centre of the University of North Carolina that works in low and middle-income countries, mainly in Sub-Saharan Africa, to increase capacity and collect, analyse and use data for improving health programmes and policies.

Discussions and debates on issues of data governance remain fragmented and compartmentalised, possibly by design to maintain the current digital health data order.55 For example, corporations and organisations could still extract health data through secondary data infrastructures and mechanisms, therefore bypassing existing laws and regulations.

Let us break things and the high cost of failure!

Mark Zuckerberg, the CEO of Meta, is quoted as saying, ‘Move fast and break things. Unless you are breaking stuff, you are not moving first enough.’56 While this may seem like the hyperbole of a billionaire, this mentality is endemic to digital innovation. Big tech companies, with their unlimited budgets, can afford to absorb smaller tech start-ups even with innovations that have not been developed and defined.57 This acquisition is premised not on making a profit out of the start-ups but rather on protecting their market dominance by monopolising the market and exterminating competitive products and services. Moreover, Microsoft and Alphabet are aggressively poaching tech talent from global south countries such as Kenya, Uganda, Nigeria and South Africa, leaving small tech start-ups understaffed, most of which cannot compete for the huge salary and benefit packages often threefold what is offered locally.58

While the African tech start-up ecosystem has been credited as one of the fastest-growing in the world, it has also seen a significant number of closures and shutdowns. This is due to the experimental nature of these tech start-up businesses and limited funding opportunities, which give an unfair advantage to the big tech conglomerates.59 For instance, the increased experimental approach to digital health that is endemic in digital health application developments in East Africa has detracted from efforts to realise universal health coverage and fragment the healthcare landscape in the region.60

From generalisations to empirical studies of national contexts

A recent commentary in the Lancet argued that with ‘the world’s largest burden of disease and the most severe shortage of healthcare workers’, African countries could benefit most from the proliferation of digital health solutions.62 Generalisations about multi-layered and multi-dimensional contexts appear regularly in both the mainstream press and in development documents. Continents like Africa are diverse, with myriad ranges of experiences. There is a need for empirical studies that appreciate this diversity and look at what works where and in what contexts.

White elephants: do we need an app for this?

There is a replication of digital health applications ranging from payment to health insurance and telehealth services everywhere. During the pandemic, we have also seen the deployment of numerous COVID-19 applications redundantly in the same settings.63 64 In a typical colonialist fashion of divide and rule, these practices merely fragment and weaken existing health systems in places and times where they are needed the most. The approach of an app for anything further seeks to individualise health in global south settings, such as in Africa. We argue that the digitisation of health should be pursued to solve existing problems, not to look for problems.

From digital sovereignty towards the pursuit of proper regulation

While digital sovereignty is seen as an approach to digital health regulation, in the global south, it might be more beneficial to have harmonised regulatory structures of resistance for pursuing health justice and solidarity. Applying theory-in-practice such as collectivism and conviviality, global south countries can work together in conceptualising and developing health data regulatory infrastructures that can be shared across nations. Structures and modes of resisting digital health coloniality can also be harmonised for a more effective and inclusive decolonial approach.