The following is an edited transcript of an interview which took place with Marion Laurence.
What is data-driven peacekeeping? What are the benefits and drawbacks?
I would describe it as the tools and practices intended to improve the quantity and quality of data that is available to peacekeepers, and to those who fund, authorize, and staff peacekeeping missions. There is a lot of focus, and rightly so, on big data, statistical analysis, and quantitative analysis. At the same time, data-driven peacekeeping also covers other types of data, and it includes qualitative analysis. These different types of data and data analysis each have comparative advantages and should be seen as complementary.
Data-driven peacekeeping offers enhanced situational awareness and allows for the development of better tools to evaluate the effectiveness of peace operations. You can build better metrics for holding personnel accountable, for example. That said, there are still drawbacks and challenges that have to be addressed. This includes concerns over data bias, privacy, and confidentiality—particularly if the data that is being gathered is sensitive. There are also political sensitivities around data collection. Some countries that contribute troops are a bit wary of some of the new data collection practices, and what data will be used for. Data has its limitations as well. No amount of data, on its own, can overcome deficiencies of political will or convince people who are unwilling to act even when solid data is available.
When considering the efficacy of predictive analytics, there is definitely potential, but there are also limited resources available, and that is a major challenge for peace operations more generally. So better prediction is great, but there is the bigger question of what gets done with those predictions. For instance, do we have the resources and inclination to act, based on the data, to have a meaningful impact? Since 2015, the UN peacekeeping budget has shrunk by about 21%. Member states are putting significant pressure on the budget, and there have been cuts to spending on personnel. These technologies have a lot of great hypothetical uses, but there is still that capacity to act that is needed.
Nonetheless, data-driven peacekeeping has already seen some successes and innovations. For instance, there are some compelling data from Darfur that shows gathering data about local community tensions ended up being quite a good predictor of attacks on civilians in particular areas. This is promising. It is a clear example that data could be used to enhance performance. The key for the UN is to capture those moments of success and feed that into a broader process of organizational learning. We can collate and share best practices, while also being sensitive to the local context. What works in one place may not work somewhere else. This is where we can add nuance to our analysis and bring in contextual qualitative data to maximize impact as much as possible.
How can Canada contribute to data-driven peacekeeping?
Efforts to improve the effectiveness and performance of peace operations take many forms, including support for things like the Action for Peacekeeping Initiative, which includes a push to have better performance data available. A comprehensive assessment system is also being developed, which Canada supports. This was to be fully rolled out by summer 2020, but as with so many things, the pandemic seems to have thrown the timelines off. The Comprehensive Planning and Performance Assessment System (CPAS) is now being used in eight out of twelve missions and is supposed to be integrated into all missions by the end of this year. CPAS allows missions to systematically assess their operating environment, identify what influence they are trying to have, and assess progress towards goals using data and analysis.
A lot of Canada’s work on this front has also been indirect, through other initiatives. The Vancouver Principles and Elsie Initiative are both linked to data-driven peacekeeping. The former calls for changes related to monitoring, reporting, early warning, and prevention to help stop the recruitment and use of child soldiers. Data is critical for all of that. The Elsie Initiative is focused on enhancing the meaningful engagement of women in Peace Operations. Barrier assessments have been ongoing in several countries, including Canada, and this involves gathering data that can help us better understand obstacles that make it harder for uniformed women to be fully represented and engaged in peace operations. These contributions are very useful, and they show that data-driven peacekeeping is not a standalone issue. It is a cross-cutting theme that is tied to many other ongoing reform initiatives. It is critically important to think about how these pieces fit together and how the lens of data-driven peacekeeping can be applied to, and complement, other thematic areas.
How will multinational peacekeeping operations approach cyber integration and interoperability? Could challenges linked to the accessibility of technology and digital literacy create tensions among those contributing, or deepen inequality among UN members?
The diverse multilateral character of UN Peace Operations is a great strength. creates challenges related to interoperability on all fronts, not just cyber. Troop and police-contributing countries are going to be arriving with very different skill sets and will be familiar with different types of technology. There will be varying levels of data literacy. The UN’s approach to addressing these types of challenges is to create standardized pre-deployment training. There is also induction training when peacekeepers start a mission, followed by ongoing training as needed to set common benchmarks, mitigate differences, and overcome discrepancies. If there is a necessity for specialized skills and a familiarity with very specific types of technology, I think this raises questions about longer-term capacity building. That is not something you can address simply in pre-deployment training.
Since 2015, the UN has been developing a Triangular Partnership Programme, which helps troop-contributing nations develop new skills over the long term in partnership with the UN and other member states. The idea is to assist troop contributors in preparation for deployments—things like capacity building in specific areas, like medical, and engineering capabilities. Gearing some of this work toward advanced data gathering and analysis could help bridge some of those potential gaps.
What strategies could be used to hold AI accountable and ensure it is as unbiased as possible when analyzing data?
This is a really important question, and it applies to all types of data collection and analysis. One good way to address this is to use intersectional analysis, what the government of Canada would call GBA Plus, very early on in the process. We need to examine how algorithms and training processes are being developed, as well as how we are thinking about what data needs to be collected in the first place. We have to ensure that sources of bias are identified early on and that creative steps are taken to minimize these biases as much as possible. An intersectional lens would be useful in terms of identifying bias and limiting it from the start. Such a process would allow us to weed out some of the potentially harmful side effects that these data-driven processes might have on specific communities.
An important thing to remember is that there should be complementarity between quantitative data analysis and a more grounded, qualitative understanding of particular settings. We should do more thinking about how to use those together in ways that can enhance peacekeeping effectiveness in all areas. Data-driven peacekeeping is not its own silo. It cuts across many thematic areas and initiatives in the world of peacekeeping.
Dr. Marion Laurence completed her Doctorate in Political Science at the University of Toronto, where she specialized in International Relations. She holds a Master of Arts in Political Studies from Queen’s University and a Bachelor of Arts (Honours) in History and Political Science from Dalhousie University. Her research has been supported by grants and fellowships from the Social Sciences and Humanities Research Council of Canada, the Ontario government, the University of Toronto, and Global Affairs Canada. She is a research associate with the Centre for International Policy Studies at the University of Ottawa.