[Read more session reports and updates from the 14th Internet Governance Forum]
The discussions on data governance that took place 5 or 10 years ago have evolved significantly. Today, we are witnessing fast developments in artificial intelligence (AI), which are changing the data landscape. Business models are evolving from data-based models, relying mostly on personal data, to AI data-based models, relying on data used in the development of AI systems. Data governance is following suit.
The IGF 2019’s data governance track, one of this year’s three main themes, concluded with a wrap-up session on current and emerging thoughts on data governance.
This track was split into six issues:
- Cross-border data: with rapporteur Mr Miguel Candia (Paraguay Ministry of Foreign Affairs)
- Jurisdictional and sovereignty issues: with rapporteur Mr Betrand de la Chapelle (Internet & Jurisdiction Policy Network)
- Data protection frameworks: with rapporteur Ms Stephanie Li (Ambassador for Netmission.asia)
- Data and sustainable development: with rapporteur Mr Wai Min Kwok (UN Department of Economic and Social Affairs)
- Human rights and Internet ethics: with rapporteurs Ms Concettina Cassa (Agency for Digital Italy), and Ms Julia Grandfield (Young Diplomats of Canada)
- Governance and ethics of AI and algorithms: merged with the previous topic for the purpose of the wrap-up session.
A long-standing issues was the dichotomy between the rights of end-users and the needs and interests of companies. Although we often expect companies to take on a more proactive role in upholding human rights, companies’ bottom line is to generate profits. Somehow, we need to reconcile these rights and needs if we want to develop a proper framework of international standards, regulations, and legislation.
One of the main challenges is the regulatory gap between developed and developing countries. When regulation is more advanced in developed countries, developing countries often have no choice but to go along with these rules, even if in principle they might not agree with them.
Jurisdictional and sovereignty issues
Data governance discussions need more clarity on which aspects are being addressed. For instance, we often fail to distinguish between the different layers of the Internet infrastructure (telecommunications, standards, and content and applications), and assume that data discussions are related to any of these. Most often, however, discussions are about data as it relates to the content layer.
When it comes to misinformation and data, while we can have very divergent views on whether restricting content is the most appropriate strategy, the main issue is the misuse of data for targeting users during the democratic process.
There is free flow of data and data protection across jurisdictions. And then there is data sharing. Data can be shared by individuals (for instance, users wanting to move their data to another service provider), and by companies (for economic gain). There is also a public interest element to sharing, such as traffic data which can be used by authorities for traffic management. In brief, these three dimensions need to be discussed together.
And finally, while we can spend time discussing which is the appropriate venue for data governance, this should not derail us from the core of the debate. There are many existing platforms which work more-or-less well for the purpose they have been set up for. For data governance, however, there is no one forum which actually brings all the actors together.
Data protection frameworks
Every country has its own needs for data protection, making it very difficult for countries to agree on a harmonised framework. With so many divergent needs and views, a one-size-fits-all global data protection framework is very difficult to achieve.
A bottom-up solution, which taps into the dynamics that civil society can create, may be productive. One example is Kenya’s approach, which it has been developing in collaboration with the Council of Europe.
Data and sustainable development
The issue of missing data is still predominant. So are issues related to how to generate and assess data. When we discuss data governance models, we refer to policies, standards, protocols, and open data. But the question remains the same: how do we apply and monitor the implementation of data-related solutions?
Human rights and Internet ethics
The current plurality of frameworks makes it difficult to understand what they have in common. If we look at AI strategies, a starting point could be agreeing on a definition for AI. And yet, current definitions are fluid. UNESCO’s ongoing work on developing a human rights-based AI framework may resolve some of the issues. This will be the first set of standards on the subject to come from a UN body.
Among all of the issues mentioned during the wrap-up session, discussions on data sharing are most likely to increase, as a participant also confirmed. Future discussions on how data should be shared will be important for countries’ digital industralisation, especially as AI developments continue to advance at such a fast pace.
By Stephanie Borg Psaila