Modelling Neo-Colonial Extractivism
“The concept of data colonialism ... highlights the reconfiguration of human life around the maximisation of data collection for profit” (Couldry and Mejias, 2019)
By: Yasmine Boudiaf
Introduction
This essay aims to demonstrate a way of processing and presenting complex social phenomena in a way that can be universally understood on a conceptual level and proposes mathematical modelling to describe the conditions for neo-colonial data extractivism.
Tech solutionism and soft power
Technological solutionism is a capitalist practice that upholds the ideology that any social problem can be solved by a technological intervention. For technology companies in particular, this cultivates a positive moral image for the business (Nachtwey and Seidl, 2020). Part of the rhetoric of tech solutionism is the implied virtue of progress; that change is good, and the old ways of doing things are intellectually and morally inferior. This philosophy mirrors the traditional colonial expansion of states in the 19th century, which assigned virtue to the economic expansion into colonies and disregarded the sovereignty of native populations (Headrick, 1979).
Soft power is the colonial practice of a state exerting influence over another state through ostensibly benevolent interventions. The effect of this is the target state’s positive attitude towards the colonial state. That also comes with better trade relations, such as favourable contracts for natural resource extraction of rare minerals (Vuving, 2009). The goal of soft power is to win the hearts and minds of the target state’s population. The native population must remain passive, or better yet, be convinced that what is taking place is a good thing. Creating a positive image of the extractive state involves psychological manipulation; influencing the popular imagination. This is a living legacy of colonialism, described by Anibal Quijano as coloniality (Quijano, 2016) and manifests in many ways, including modern extractivism. By controlling the narrative, restricting access to information and emotionally pacifying the population, the extractive state can carry out their actions with the least amount of resistance.
“Soft power leads to two main categories of outcomes (cognitive and behavioral) at micro, meso and macro levels, namely individual valence, group and public opinion, as well as individual behavior, institutional policy and foreign policy” (Davis and Ji, 2016).
It is the capitalist endeavour of perpetual economic growth that drives extractivism. Within that, there is neo-colonial extractivism, where historical colonial powers continue to extract resources from ex-colonised countries (Acosta, 2017). The definition of neo-colonial extractivism can be expanded to include any valuable resource a target group of people controls, and to involve any entity, a state or otherwise, that seeks to extract that resource. Instead of a colonial state, the extractor can be a technology company, and instead of rare minerals, the target valuable resource can be personal health data.
There are proven conditions for successful soft-power interventions demonstrated throughout history (Rawnsley, 2016). Many of these conditions can also apply to a country for the purposes of data extractivism:
- Pre-existing power dynamic to leverage
- Subjugation of native people to external powers
- Extractor’s claim to to superior rationality and knowledge as justification
- Reductionist approach towards target population
- Money or service given in exchange for the resource is many orders of magnitude lower in value that what the resource will go on to be sold for
- The practice benefits economies and populations outside of the place where extraction is taking pace more than it does the population where extraction is taking place
- Individual, population or state sovereignty over the valuable material concerned is not upheld
- International laws and conventions are selectively employed
- Multi-step extraction process from raw material to final product with an increase in value with each step
- Employment of cultural soft-power practices
- Rationality, progress, modernisation or altruism as framing arguments to justify extraction
- Conversion of natural resource into capital
- Obfuscation and undemocratic mechanisms
Personal health data
Technological solutionism can be applied to healthcare, particularly where it concerns patient data. The argument often used is that datafication, including use of machine learning, will result in better health outcomes for the patient, although this has not been proven true in all cases (La Cava et al., 2019). Owing to the digitisation of patient health records worldwide, there is opportunity for health tech companies to commercialise that data. The worldwide trade in personal data is highly lucrative and worth hundreds of billions of dollars (Lieshout, 2015). Personal data offers insights into human behaviour and vital statistics. This is valuable information to a company or a political power wanting to sell a product or idea to a particular group of people.
This practice raises ethical concerns around privacy, security, consent and agency. International regulations such as the GDPR in the EU (Voigt and von dem Bussche, 2017) are designed to protect an individual’s right to privacy and give agency to the individual over the use of their personal data. Despite such regulation, individuals unknowingly hand over personal health data through normal social and economic activities (Paul et al., 2015). Kuntsman et al. address the difficulty in opting out of health data sharing and aggregation (Kuntsman et al., 2019).
Only 66% of countries worldwide have data protection regulations in place (“UNCTAD | Data Protection and Privacy Legislation Worldwide,” 2020). Countries that do not enforce such policies are good candidates for data extractivism. Similar to natural resources, personal health data can be mined, extracted, used and sold. The data workflow is surprisingly similar to mineral extraction:
- Identify valuable data.
- Convince gatekeepers to grant access to the data or take the data.
- Mine or harvest the data.
- Clean the data.
- Analyse the data.
- Use or sell the data.
With each of these steps, consent or awareness from individuals within the target population concerned is not necessary, in the same way that consent and awareness is not necessary in neo-colonial natural resource extraction (Barocas and Nissenbaum, 2013).
Babylon Health: a case study
A current example of a tech company displaying neo-colonial extractive and soft power tactics in order to harvest personal health data Babylon Health, a British health-tech startup.
Babylon Health was contracted by the UK’s National Health Service to offer GP At Hand to select patients; an app that claims to use artificial intelligence to perform a patient assessment, at the same rate of accuracy or better than a human doctor (“GP At Hand,” 2020). Despite concerns from doctors and patients about these claims (“Patients and GPs gather for protest against GP at Hand,” 2018), and wider concerns around how the company came to win such a contact in the first place (“Conflict of interest questions over Cummings’ job at health tech firm,” 2019), Babylon Health successfully partnered with the governments of Rwanda, Canada and the US to roll out similar apps.
Its behaviour fits many of the conditions for neo-colonial extractive practice as it concerns personal health data, and employs soft-power tactics:
Modelling behaviour
Boaventura de Suosa Santos argues for the importance of epistemological justice in global justice (Santos, 2007). He points out the inherent power that is bestowed on those who control the framing of knowledge. This is also true for semantics and the use of language as it concerns definitions and context. Actions can be considered justifiable if the perpetrators are the ones defining justice. If there was a way to detach the epistemological and linguistic properties of social phenomena, perhaps a conceptual sense of the power dynamics at play can emerge, that are not tainted by the actual power dynamics that may exist (as a result of historical colonialism, for example). Such altering of the politics of representation requires an intervention, that, as described by Marisol de la Cadena as is effective “without canceling—(the practice of) categories, concepts, or analytics that may overpower, perhaps even kidnap the situation that is up for description” (de la Cadena, 2017).
The intervention proposed is mathematical abstraction. The purpose of abstraction is to move away from a semantic description to one that is symbolic, yet still be relevant to real-life mechanisms. Describing the actors and power dynamics involved in neo-colonial extractive practice semantically is inherently limiting, owing to the origins of the descriptors used and the position of the person or entity employing them. It is also not possible to capture all detail in a complex system in a snapshot description.
With category theory, related semantic elements are given symbols in a higher level of abstraction than the detailed elements. The purpose of category theory is:
"1. To provide a language for making precise statements about mathematical concepts, and a system for making clear arguments about them.
2. To idealise mathematical concepts so that a diverse range of mathematical notions may be compared and studied simultaneously by focusing only on relevant features common to all of them" (Eugenia L. Cheng, 2000).
The relationships between elements can also themselves be given symbols and participate in the same system as the elements, in a higher dimension, in what is described as higher-dimensional category theory. Those relationships in turn can have relationships with each other. Interactions between these relationships have rules that follow a prescribed logic. A change in the way higher-dimensions are arranged will influence the underlying architecture of elements in lower dimensions.
String diagrams (topological graphs) are a useful and visually captivating way to describe such relationships. Using algebraic substitution the agents, decisions and consequences can be displayed and their relationships can be manipulated. This notation is usually used in quantum physics and computing, but has been used by Hedges et al. in game theory to describe economic decisions:
Discussion and conclusion
The methods used to describe social mechanisms are themselves important, as they can shape the narrative around the issues concerned moving forward (de la Cadena, 2017). The practice of mathematics has often been scruitinised as a tool that abstracts away from the issues concerned. There is a tendency to take emotion and human relevance out of the equation.
While this has been the case historically, such detachment from detail can allow for a more conceptual understanding of the phenomena described. Our innate familiarity with topological structures that resemble those found in nature should be further explored.
Mathematics as a method is designed to be universally understood, with potential to act as a counter-tool to the obfuscation that is necessary for soft-power and neo-colonial extractive practices to be successful. Although it is co-opted by tech-solutionism, it is not inherently partial, and there is opportunity for anyone to take the practice in any direction.
A single, situated agent constructing such a model will inevitably produce an incomplete description. A proposed way to overcome the limitations of partiality, detachment and social distance is to open up the exercise to involve the concerned actors. In this case, it would be each of the underlying elements in the system (for example, getting patients whose health data is being extracted to help construct opetopes).
Another limitation is that such models are static and inflexible and so do not reflect changes in the underlying semantic infrastructure (such as change of agents concerned or the effects of time). A flexible, dynamic model, which exists online and can receive relevant information that influences its architecture would be a better reflection of the nature of the phenomena it is modelling.
Author's note
This essay describes events from a geographical and social distance, and with that detachment comes an inherent reductionism that may resemble the negative behaviours described, or risk reenacting a type of colonial position, described by Nishat Awan as “engaging with places at a distance” (Awan, 2016). To reduce this possibility, the focus of this inquiry is on the descriptors used when investigating soft-power practices, as opposed to offering an opinion on any group of people or countries.
The author’s concern is around the use of semantics, symbols and mathematical modelling as processes of enquiry. It is proposed that:
- The definition of neo-colonial extractivism be expanded to include personal data capture through soft-power mechanisms.
- The practice of abstract mathematical modelling when used to describe social phenomena should a) where possible, include contributions from the actors it concerns to produce a collective artefact that represents the social phenomena truthfully and b) be constructed to physically evolve over time to reflect structural changes caused by shifts in power distribution, changes in agents and time.
References
Acosta, A., 2017. Post-Extractivism: From Discourse to Practice—Reflections for Action. Int. Dev. Policy Rev. Int. Polit. Dév. 9, 77–101. https://doi.org/10.4000/poldev.2356
Awan, N., 2016. Digital Narratives and Witnessing: The Ethics of Engaging with Places at a Distance. GeoHumanities 2, 311–330. https://doi.org/10.1080/2373566X.2016.1234940
Barocas, S., Nissenbaum, H., 2013. Big data’s end run around anonymity and consent. Priv. Big Data Public Good Framew. Engagem. 44–75. https://doi.org/10.1017/CBO9781107590205.004
Conflict of interest questions over Cummings’ job at health tech firm [WWW Document], 2019. . Bur. Investig. Journal. URL https://www.thebureauinvestigates.com/stories/2019-10-11/conflict-of-interest-questions-over-dominic-cummings-job-at-health-tech-firm-babylon (accessed 5.18.20).
Couldry, N., Mejias, U.A., 2019. Making data colonialism liveable: how might data’s social order be regulated? Internet Policy Rev. 8.
Davis, R., Ji, L. i, 2016. Challenges of a big data approach in mapping soft power. Routledge Handbooks Online. https://doi.org/10.4324/9781315671185.ch7
de la Cadena, M., 2017. Matters of method; Or, why method matters toward a not only colonial anthropology. HAU J. Ethnogr. Theory 7, 1–10. https://doi.org/10.14318/hau7.2.002Eugenia L. Cheng, 2000
Eugenia L. Cheng, 2000. Category Theory, http://eugeniacheng.com/wp-content/uploads/2017/02/cheng-architecture.pdf
GP At Hand [WWW Document], 2020. . GP Hand. URL https://www.gpathand.nhs.uk/ (accessed 5.18.20).
Headrick, D.R., 1979. The Tools of Imperialism: Technology and the Expansion of European Colonial Empires in the Nineteenth Century. J. Mod. Hist. 51, 231–263. https://doi.org/10.1086/241899
Hedges, J., Shprits, E., Winschel, V., Zahn, P., 2016. Compositionality and String Diagrams for Game Theory. ArXiv160406061 Cs.
Kuntsman, A., Miyake, E., Martin, S., 2019. Re-thinking Digital Health: Data, Appisation and the (im)possibility of ‘Opting out.’ Digit. Health 5. https://doi.org/10.1177/2055207619880671
La Cava, W., Bauer, C., Moore, J.H., Pendergrass, S.A., 2019. Interpretation of machine learning predictions for patient outcomes in electronic health records. ArXiv190312074 Cs Stat.
Leinster, T., 2003. Higher Operads, Higher Categories. arXiv:math/0305049.
Lieshout, M., 2015. The Value of Personal Data. Presented at the IFIP Advances in Information and Communication Technology, pp. 26–38. https://doi.org/10.1007/978-3-319-18621-4_3
Marsden, D., 2014. Category Theory Using String Diagrams. ArXiv14017220 Cs Math.
Nachtwey, O., Seidl, T., 2020. The Solutionist Ethic and the Spirit of Digital Capitalism (preprint). SocArXiv. https://doi.org/10.31235/osf.io/sgjzq
Patients and GPs gather for protest against GP at Hand [WWW Document], 2018. . Digit. Health. URL https://www.digitalhealth.net/2018/03/gp-at-hand-protests-east-london/ (accessed 5.12.20).
Paul, M.J., Sarker, A., Brownstein, J.S., Nikfarjam, A., Scotch, M., Smith, K.L., Gonzalez, G., 2015. Social media mining for public health monitoring and surveillance, in: Biocomputing 2016. WORLD SCIENTIFIC, pp. 468–479. https://doi.org/10.1142/9789814749411_0043
Quijano, A., 2016. Coloniality of Power and Eurocentrism in Latin America: Int. Sociol. https://doi.org/10.1177/0268580900015002005
Rawnsley, G., 2016. The Routledge Handbook of Soft Power.
Santos, B. de S., 2007. Cognitive Justice in a Global World: Prudent Knowledges for a Decent Life. Lexington Books.
UNCTAD | Data Protection and Privacy Legislation Worldwide [WWW Document], 2020. URL https://unctad.org/en/Pages/DTL/STI_and_ICTs/ICT4D-Legislation/eCom-Data-Protection-Laws.aspx (accessed 5.13.20).
Voigt, P., von dem Bussche, A., 2017. Scope of Application of the GDPR, in: Voigt, P., von dem Bussche, A. (Eds.), The EU General Data Protection Regulation (GDPR): A Practical Guide. Springer International Publishing, Cham, pp. 9–30. https://doi.org/10.1007/978-3-319-57959-7_2
Vuving, A., 2009. How Soft Power Works. SSRN Electron. J. https://doi.org/10.2139/ssrn.1466220