HumanitarAI delves into the dynamic intersection of datafication, AI, and social media, exploring how these tools are reshaping the way we approach humanitarian efforts and communication for development.
 
Why Data is not always the solution

Why Data is not always the solution

The Pitfalls of Datafication

Data or Big Data is increasingly being used for the development cooperation sector to measure and analyze global problems, while simultaneously suggesting that within this data lies the solution. The World Bank specifically refers to it as a ‘data revolution’ where they collect, analyse and use big data to monitor and achieve the Sustainable Development Goals and refer to it as Data Driven Development (World Bank Group, 2018). As per a previous blogpost, the importance of Big Data has become a Global Power Struggle. Not only is there the issue of Data Governance across different regions but also a market for data and a race to collect as much as possible and be the first to do so (Jagtiani, 2023). This can be illustrated by increasing investments by the government into the China Big Data Centre on issues such as AI in order to ‘outpace the US and other Western nations in setting norms and standards’. (Cai & Zheng, 2023) This illustrates the importance of critically examining the use of data in humanitarian contexts and beyond. Mary Poovey (1998) demonstrates how across history, all the way back to the 16th century, humanity has privileged numbers and statistics as a universal and absolute truth in the establishment of modern facts. It is not new for us to rely on numerical representations or ‘data’ as the most accurate reflection of reality. However, that doesnt mean that it is. And in the age of the data revolution it is our moral obligation to assess the way to which we are representing and interacting with the world. 

Data is not the only Truth

Not only is the harvesting of data questionable due to issue of consent of privacy rights as mentioned in my first blog post, but it also doesnt allow for participatory methods. A common example is the use of Satellite Imagery for measuring and assessing changes in the SDGs. Earth Observations and Geospatial Data have become an increasingly affordable method of assessingair and water quality, mapping land, development, and property rights as well as diseases and natural hazards (World Bank, 2017). China just launched its own Satellite called SDGSAT-1 in 2021 for this very purpose, countering Sentinal-1 and 2 from Copernicus (USA) and the European Space Agency whos data has been traditionally use by large intergovernmental organisations and as default global data sets further exemplifying geopolitical contentions and their potential intensification. Satellite imagery can assess vegetation cover which is the percentage of soil covered by green vegetation however local and national actors often complain that it provides inaccurate results when there is a variety of species in regard to the vegetation and soil classification at agricultural sites (Schönau, 1987, p. 13). While that journal article was published in 1987, the debate still exists and in various formats, as data has increased in volume and methods of analysis in variety. While this may seem like a minute detail a part of a small sub-indicator of an SDG it can result in differences in support that regions and communities of the world are given or are paid attention to. 

Mark Twain is famous for a quote that says “It ain’t what you dont know that gets you in trouble. Its what you know for sure that just aint so.”. In the same vein, the way that experts believe this data to be the truth, it is actually just an opinion; one perspective. Along with the assessment on vegetation comes a value-added judgment. For environmental organisations high vegetation cover is good, while for socioeconomic reasons land is required for other resources. Finding a balance just between the value judgments of the data itself is as unlikely to be resolved and determining a single ‘truth’. Furthermore, this data is rarely evaluated by local or national stakeholders for its reliability and accuracy. While other countries have sophisticated space agencies and satellites, Western data continues to dominate the market. It is considered to be the highest quality, is used in all of the largest tools and methodologies such as Google Earth Engines and GIS plugins as well as the default data for monitoring and measurement on all UN agencies. 

Participatory methods include on-the-ground surveys, and numerous intersectoral workshops between local and national experts and political. All of this not only costs time but also money, which is why it is rarely incorporated alongside big data for humanitarian considerations. When Turkiye was given a multi-million euro fund to perform such a participatory method for the Upper Sakarya Water Basin they developed a Decision Support System for the entire country which estimated 49% of the country’s land affected by drought while the default data based on Western Satelittes and without validating methods would have indicator below 20% (UNCCD Dashboard, 2023). While this particular case study is extremely specific, the numerical implications of not using participatory methods are extremely alarming. In this way, large humanitarian agencies are using data in such a way that further removes the individuals affected by these issues and indirectly reduces their own voice. As such, Read et al., (2016, p. 12) accurately refer to this ‘Data Hubris’ as a system that replicates and ‘reinforces existing power holders’ making it far ‘less than emancipatory’. 

Participatory Methods

Are we reducing people to numbers on a screen? Some might saw we have always assigned humans certain value sets based on their perceived contribution to society such as a man able to fight in battle or a woman able to bare children. Recently we have started to collect data on human beings in such a way that thez are “assessed, profiled, categorised, and “scored” according to data assamblages” (Hintz, 2021, p. 3549). Not only is there a greater recognition that people are being subjected to extensive data analysis with significant affects but it is also evidence that this is being done “without much understand, voice, and influence” on the part of citizens (Hintz, 2021, p. 3555). Therefore it a pejorative of the agencies collecting and analysing this data to also include the voices of whom that data is being collected from and ultimately effects. 

Admittedly this can be be a complicated process as being a participant requires being able to understand, access and process the ongoings of dataification. Hintz (2023, p. 3550) poses relevant questions that we must all consider as barriers to entry:

“How, then, do we participate as active citizens in a society in which we are constantly assessed according to data analytics that we cannot access or engage with? 

How do we intervene into algorithmic governance processes and affect the development and management of the very data systems that increasingly organize society? 

How do we maintain and expand civic participation in a context of rapid technological and social transformation, and how do we develop new democratic processes to ensure participation, transparency, and accountability?.”

Such considerations can be adapted to the collection of data across different contexts. In the context of geospatial data, countries with higher rates of cloud cover are subject to less available and reliable data. The majority of these countries also just so happen to fall under the status of least developed states. Therefore not only to they have less access to information that could support them, they are also less likely to have the resources to make up for this gap in technology. Another example would be Small Island Developing States where some Earth Observations working at scales of 1km by 1km make up just a couple of pixels on a screen. If we continue to rely on big data we risk further marginalised communities that are already those most affected by Global Issues. 

This does not even take into consideration the multitude of issues regarding the collection of data based on epistemological norms vs. understanding and analyzing data across languages and cultural standards which could fill multiple volumes. Ultimately data is meant to serve the goals of humanity such as the SDGs rather than stand in the way of them. In lamens terms, different data will have different values depending on methodologies or context and needs which reflect the diverse and messy reality inherent to international development. It is our own hubris that would assume we could reduce the most complex issues in the word to single numerical values. At the end of the day, it is humans who make these issues so complex, so we might as well accept our own complexities rather than try to reduce and simplify them in the risk of misrepresenting them.

So what should we do?

In no way does this article suggest that we abandon the use of big data in addressing Global Issues. In actuality, the data revolution has provided us with more opportunities to support our communities. However, our tendency for clarity and perfectionism has led to oversimplification and distorting representations of reality. Instead of analysing data on a screen, we need to spend more time talking to others, exchanging ideas and being willing to disagree and compromise. If we spend time, money and resources on data technology, it is imperative that we spend the same on participatory methods and engaging with human beings directly affected by these issues. Too often are organisinations like the UN criticised for not making a difference and with an increase in dataification we are likely to further distance ourselves from complex realities and the people living them. 

Reflections

It may be a bias coming from someone who prioritises communication and effective stakeholder engagement in her work within development cooperation but it seems that the higher up you go, the more people are speaking a language that most dont understand. Scientists and UN veterans who squabble over the definition of land use or land management, making speeches about NDVIs (Normalized difference vegetation index) and using phrases that were most likely invented just for the organisation to be able to talk about it. I have never understood why we admire intelligence if someone is unable to make the people around them not only understand what they are saying but actually contribute to the topic at hand. The use of data, the language surrounding datatification and the algorithms and methodologies involved are not only complex, they are alienating, and not just for communications specialists. The use of Earth Observations and Geospatial Technology was new to me but it was also exciting to see how much information could be captured through these methods. What has disappointed me is the fact that national policymakers and stakeholders are never even part of the discussion on standards, norms, development of tools, or datasets. There may be representatives of Civil Society Organisations at conferences but the notion that the small-hold farmers suffering from famine and drought may have valuable insight on how we measure and address these issues does not seem to be considered ‘common sense’. Datafication could revolutionise the way we address SDGs but at the moment it is only further disenfranchising already marginalised communities. It is the people with the power and money who sit at the table and make decisions, none of whom have ever come close to suffering from the issues upon which the data is trying to help us better understand.

References

Cai , J., & Zheng, W. (2023, November 20). Why China’s new data agency shows ambitions to develop digital economy and AI. South China Morning Post. https://www.scmp.com/news/china/politics/article/3242067/china-launches-new-data-agency-ambitions-ai-and-digital-economy-soar

Hintz, A. (2021). TOWARDS CIVIC PARTICIPATION IN THE DATAFIED SOCIETY: CAN CITIZEN ASSEMBLIES DEMOCRATIZE ALGORITHMIC GOVERNANCE? AoIR Selected Papers of Internet Research. https://doi.org/10.5210/spir.v2021i0.11943

Poovey, M. (1998). A History of the Modern Fact Problems of Knowledge in the Sciences of Wealth and Society. University Of Chicago Press.

Read, R., Taithe, B., & Mac Ginty, R. (2016). Data hubris? Humanitarian information systems and the mirage of technology. Third World Quarterly, 37(8), 1314–1331. https://doi.org/10.1080/01436597.2015.1136208

Schönau, A. P. G. (1987). 7. Problems in Using Vegetation or Soil Classification in Determining Site Quality. South African Forestry Journal, 141(1), 13–18. https://doi.org/10.1080/00382167.1987.9630255

Sharinee Jagtiani. (2023, August 10). The Geopolitics of Data Governance and Digital Power Play | GJIA. Georgetown Journal of International Affairs. https://gjia.georgetown.edu/2023/08/10/the-global-cloudscape-the-geopolitics-of-data-governance-and-digital-power-play/

UNCCD Dashboard. (2023). UNCCD Data Dashboard. Data.unccd.int. https://data.unccd.int/country-overview?country=TUR

World Bank. (2017, July 23). Using Satellites to Monitor Progress toward the SDGs. World Bank. https://www.worldbank.org/en/news/feature/2017/08/23/using-satellites-to-monitor-progress-toward-the-sdgs

Selected Papers of Internet Research. https://doi.org/10.5210/spir.v2021i0.11943