What does data governance mean to you?
That’s the question we asked in our Connected Conversation with Global Voices (GV) on 30th September, exploring learning from GVs Future of Data Challenge-funded Civic Media Data Observatory on Data Narratives that gathered and analysed data-related stories from El Salvador, Brasil, India, Sudan and Turkey over the first part of this year.
Data governance from the ground up
The Civic Media Observatory method uses local researchers, invited to identify issues and stories being discussed in social and local media, and to then carry out in-depth analysis of media items about these, digging into the context, subtext and narrative frames they represent. Through this ground-up approach, the observatory surfaces and offers alternative sense-making on stories that either may not have surfaced in global awareness, or that are being viewed through a narrower set of lenses in mainstream discourse. At the same time, a team of editors working across countries, are able to draw out common narrative threads between countries, organising collected media items into a structured dataset, freely available to browse through AirTable.
Our speakers (Giovana, Ivan and Tim) introduced insights from three of cases explored in the Data Narratives observatory, reflecting on how they bring both challenge and depth to discussion of data governance:
Sudan: Governance is unstable
Regimes that we build make assumptions about what people can share and the data narratives research in Sudan found how unstable data governance is where the overall government context is similarly unstable. Governance is built in times of peace and it can be thrown out in conflict. In conflict the data too can be used by opposing parties and this is being seen in the Russian invasion of Ukraine too where, as Russia took over Ukrainian territory, they switched off local data points and switched to Russian points, so Ukrainians wanting to access data had to do so via the country invading them. Similarly in Myanmar telecom companies were forced to sell their businesses to state owned companies which resulted in the data becoming subjected to state surveillance.
In conflict there is competition for control of the internet and telecoms for messaging and surveillance but it can also be a source of revenue and empowers whoever owns the access. The research suggests such experiences will likely be seen playing out in all countries in conflict. In Sudan it was also observed that when there is no other option, the authorities bypass their own shutdowns as the Sudanese Armed Forces opened internet connections via Starlink for themselves and their allies - using it as a way to garner favour and sharing it selectively.
Brasil: Geopolitics, big tech and domestic interests
2024 was a big policy year in Brasil with elections and hosting G20. Amongst this the “Fake News Bill” (as it became known - which in itself has a narrative impact) was a piece of legislation intended to govern speech and content online (formerly called the ‘Freedom, Responsibility and Transparency on the Internet Bill). The development of the Bill started with no-one being happy but it evolved over time and by the end it was a very detailed piece of proposed legislation, that was based on analysis, sophisticated and considered to be progress. After much time and effort from many involved parties, the Bill was dropped.
GV research in Brasil found that the internet regulation debate may hold a cautionary tale told in three acts. Governments and institutions acting too hastily can harm human rights, but taking too long to regulate as new technologies rapidly evolve can be harmful too.
El Salvador: Data is physical
There is a physicality to data governance. El Salvador banned mineral mining in 2017 due to its environmental impact but are now finding bitcoin (cryptocurrency) mining is having a similar impact. High level decision makers can view data as intangible and abstract but AI has brought the environmental impact of data more into our consciousness. The data centres being built to handle AI demands are having significant local impacts on neighbours and there are narratives about this including leaving communities without water. The research in El Salvador finds that data governance issues are different if your currency is tied up with crypto currency; data governance decisions are linked to climate and environmental issues.
There are many more stories to explore from the Civic Media Observatory on Data Narratives, both through written-up narratives, and by exploring the database of media items, each accompanied by a contextual description that helps make sense of the importance and impact of the item.
Data governance from the top down
At the same time, our conversation took place in the context of the recently agreed Global Digital Compact, and with organisations who often look at data governance either as a global issue, or as a policy problem in the context of Western democracies.
Extracts from the Global Digital Compact, agreed on 22nd September 2024
“37. …responsible and interoperable data governance is essential to advance development objectives, protect human rights, foster innovation and promote economic growth….”
“38. We recognize the urgent need for strengthened data governance cooperation at all levels with the effective, equitable and meaningful participation of all countries and in consultation with relevant stakeholders to unlock the full potential of digital and emerging technologies….”
“48 … we request the Commission on Science and Technology for Development to establish a dedicated working group to engage in a comprehensive and inclusive multi-stakeholder dialogue on data governance at all levels as relevant for development. We encourage the working group to report on its progress to the General Assembly, by no later than the eighty -first session, including on follow-up recommendations towards equitable and interoperable data governance arrangements, which may include fundamental principles of data governance at all levels as relevant for development; proposals to support interoperability between national, regional and international data systems; considerations of sharing the benefits of data; and options to facilitate safe, secure and trusted data flows, including cross-border data flows as relevant for development.”
When we think about data governance, we often have in mind a fairly technical definition, such as that put forward in this datasphere literature review:
Data governance concerns the rules, processes and behaviours related to the collection, management, analysis, use, sharing and disposal of data - personal and/or non-personal.
Good data governance should both promote benefits and minimise harms at each stage of relevant data cycles.
But how can ideals of collective, democratic, participatory, deliberative and powerful data governance be promoted as global norms in a context of unstable, geographically anchored, and geopolitically conditioned governance?
This led us to our discussion where attendees were then invited to consider the following questions:
- How does data governance present in the contexts you are working in?
- What policy measures do you work on / support to improve data governance?
- How are these supported or challenged by bottom-up narratives about data and AI?
Discussion points
It was noted there is always an interest in withholding data and a question about what transparent data could look like. An example was shared about data collected in relation to dams and water flow - who holds this and who could benefit from access to the data. There is a balance of common/public good vs data as a strategic national interest which creates the geopolitical nature of data. How do we bring public voice to that?
Giovana shared some further insight to the Turkey research which had identified significant data leaks. Due to lack of access to data provided by the Government relatively cheap access was made possible by leaks (likely from within Government departments). The impact on the governance narrative was that some considered Big Tech companies had the responsibility to ensure data protection and civil society believed all data was subject to being leaked and wasn’t secure.
There is a present danger and opportunity in the global climate crisis. A UN Convention in 1998 on data sharing proposed open source solutions across the EU. At the time it was found to be practically impossible to come to a one platform solution as every country wanted the ability to control content and access. There was an identified need for transnational solutions. It was noted that currently it is international companies rather than publics doing that transnational work currently.
There was discussion about the research methodology that had been used in the Data Narratives Civic Media Observatory. Narratives may be explicit (usually with the presence of action verb) but the research work is concerned with reviewing all the texts and testing the logic and durability over a set of items. If a narrative is only found in one item then there is limited social reach and significant impact is unlikely from one item. Contestation is often happening in a proxy space (in relation to data governance) and most narratives about data governance are not about data governance (they are intersectional and the governance element is opaque). Previous non-data governance related research has uncovered implicit/underlying data governance narratives. Where possible a taxonomy (such as implicit/explicit, weaponized or neutral, bottom-up or top-down, perceived or promoted) of the narratives is included in the description of the narrative in AirTable. In addition, in the analysis of media items in the dataset there is detail regarding who is espousing the narrative, whether they have an underlying interest, what the context of the narrative’s presentation might be, whether or not the narrative is supported by fact or disinformation, etc. Some researchers interviewed hacktivists and other communities to try to see the unseen.
It was discussed how perhaps to influence change based on this research there is a need to take the narratives, reframe them and share them back into policy making discussions. It was noted this work could be a useful tool at a macro level to engage with companies and/or policy makers. Can we draw any generalities?
In the UK there is public interest in the use of data particularly in create a new AI or new research to inform policy. The privacy argument vs public interest is a big area of contention and needs data governance to balance those. This kind of balance debate hasn’t come through in the countries considered. It was suggested that this is due to a big tension of underlying assumptions about whether state regulation and control of data is benevolent or malignant. A core assumption, for example, in the European context is usually that regulatory authorities have a public interest mindset. That is far from a universal view. This should also remind us to think about how ideas we may be promoting will play out in other countries and our need to improve at contextual conversations to influence the country specific experience. Global narratives about “public good” raises the fact that a western narrative can be experienced as a negative influence/jeopardy for many national contexts.
It was queried whether anyone was aware of research on data governance narratives from international / inter-governmental processes like G20 or the UN? It was noted that, loosely related, was a publication on the commodification of trust in global AI policy discourse and Niels Ten Oever at the Critical Infrastructure Lab was doing strong academic work on the transition of the G20 presidency from India to Brasil.
Next steps
The Global Voices team are keen to continue this research.
Attendees also mused over how we can use this narratives research to influence change: what can it mean for the advocacy messages we might all develop, or the ways we practice policy development?
Giovana welcomed any email contact from anyone wanting to continue the conversation.