Peacemaking and conflict mediation: AI as a tool, human judgment as the key

Lucas Pretti: What do conflict mediators and peacekeepers need to know about AI? How is it already being used, and how should it be used?

Miloš Strugar: AI is already an integral part of conflict-resolution processes. It is used for summarising documents, translating between languages, and for sophisticated search. For a mediator, this can help them better understand people’s interests, needs, desires and perceptions, and integrate those perspectives into the conflict-resolution process.

But we should not forget that AI is essentially mathematics, patterns, data, code and algorithms. It is non-human intelligence. In conflict resolution, you need superior human intelligence. Human judgment is the most important part of conflict-resolution processes.

So AI is an assistant. AI is a tool. There are aspects of conflict resolution, negotiation and mediation that cannot be delegated to AI. Moral and ethical reasoning is one example. Do you negotiate with terrorists and war criminals if this is possibly the only way to stop violence? Imagine that one party to a conflict is using force. Is that justified, legitimate or not? Should a mediator be guided by utilitarian or categorical moral reasoning? These are questions of moral reasoning where, in my view, it is extremely difficult – if not impossible – to replace human judgment. Political philosophers are grappling with these and other moral and political dilemmas for centuries with no conclusive answer.

We have to understand the limitations of AI while using its advantages to the maximum to assist our work.

Tanja Strugar: I would like to add something that everyone should keep in mind when using AI: the data behind the AI determines a lot of what the AI puts out. I don’t think enough people are aware of how important this is. Many datasets are primarily Western-focused, which limits their real-world application and relevance to conflict zones outside the West.

If you are dealing with conflicts between cultures in regions beyond the West, you cannot assume that an AI system will have a dataset that can objectively or adequately analyse the situation and provide the insights you need.

Lucas Pretti: Some research has tried to identify which cultural perspectives different models tend to reproduce. How does the course address this?

Tanja Strugar: The course is not primarily about individual LLMs, although they will be covered. Our focus is on developing the skills to use them effectively. We want to equip people to identify where an application can be useful, how it can be used and what its limitations are.

We will look at applications such as conflict mapping, stakeholder analysis and cultural narratives. AI will continue developing and its coverage will continue expanding. So rather than teaching people something that might become obsolete, we want them to reach a level where, regardless of what happens in the AI landscape, they know how to use these technologies and how to get the most out of them without putting anything at risk.

Miloš Strugar: The important question behind this discussion is: can AI contribute to greater equality between nations, communities and people?

Before developing this course, Tanja and I spoke with practitioners currently working in the field. We asked them whether they were using AI, how they were using it, whether it was useful and where they would like to see AI used more. One of the questions that emerged very clearly was whether AI is contributing to equality specifically in conflict-resolution processes.

Our moral views and world views are shaped by our culture. When we talk about values, we have to remember that different cultures have different values, and same values may have different meaning in different cultures. Take the word “peace”. In Europe, peace may be understood in terms of order and the rule of law. In Japan, China or India, it can be associated more with harmony, tranquillity and cosmic order. Peace can mean the absence of violence, or it can mean cooperation.

So if an AI system tells me, “This could lead to peace,” without explaining what it means by “peace”, I cannot necessarily use that answer. This is a very important aspect of responsible AI in conflict resolution.

Lucas Pretti: Conflux has been around for about ten years. What has changed during that time, both in the broader context and in the way you work?

Tanja Strugar: I would say the pivotal moment for us was the beginning of COVID-19, when we transitioned into online courses. We have always been very practice- and skills-focused. We cover theory only to the extent that it explains the context in which a skill can be used. We want participants to leave with the feeling that they have achieved something that will actually help them in their day-to-day work.

Moving online meant investing more in materials and methods that could create that immersive experience even when people were sitting in front of a screen.

At the same time, our topics became more diverse. We started with a relatively narrow focus on negotiations, conflict mediation and conflict prevention. As we grew, connected with more experts and received feedback from participants, we began to see connections between different areas of work.

Miloš Strugar: Another pivotal moment was the changing geopolitical situation. We are at a turning point that makes mediation much more challenging. Stakeholders are changing, and it is more difficult to find common ground.

In response, we introduced another form of activity: non-formal dialogues. We bring people from opposing sides or different parts of the political spectrum together, not necessarily for mediation or track-two diplomacy, but simply to better understand each other. We have been doing this regularly for several years, bringing Europeans and Iranians together, Americans and Russians together, and other groups with a relatively modest objective: to improve understanding and open a dialogue.

We also provide bespoke courses. For example, for a mission in South Sudan, we developed courses around negotiation and mediation of intercommunal conflicts based on the participants’ actual situations. Rather than using a made-up scenario, we worked with the real issues they were dealing with every day.

Lucas Pretti: Are conflict-resolution processes happening online today? Who are the people working on peace negotiations?

Miloš Strugar: There are obvious limitations for real substantive negotiations online. Negotiators certainly use online platforms, and I even give lectures about how to negotiate online because there are some practical advantages to it. But face-to-face contact is irreplaceable for peace negotiations.

Very often, the cultural aspect of a conflict is neglected. You have objective factors – interests, needs and the public positions of the parties – and these can often be negotiated if there is room for negotiation. But there are also subjective factors related to culture.

In some cultures, the outcome is the key: you have a problem, so the question is how to resolve it. But in many other cultures, the relationship is the key. You focus on building the relationship, and the outcome comes as a result of that relationship, trust and confidence.

For some, it may be possible to negotiate online. But where relationships are central, it becomes much more difficult.

Tanja Strugar: In our programmes, we normally address senior- to mid-level staff members in international organisations, field missions and operations. The UN tends to make up the majority of our attendees, but we always have an open door. We have had people from the private sector, governments, non-government organisations and academia join our programmes as well.

The topics we cover are applicable beyond organisations such as the UN. For the most part, we are talking about stakeholders and decision-makers working within conflicts and highly sensitive environments, as well as people who need practical skills that they can immediately apply in their work.

Lucas Pretti: How do you see this course connecting to PeaceTech?

Tanja Strugar: There is definitely a convergence between the intention of this course and PeaceTech, especially the mission of the PeaceTech Alliance, particularly when it comes to making digital tools more familiar and usable for peacebuilders and peacekeepers.

We are approaching it from a different angle. Rather than looking primarily at the technology itself, we are looking at the human first. When you have a practitioner who understands both the technology and its limitations, you can create a much better fit between the tool and the work being done on the ground.

Miloš Strugar: I think we are all moving in the same direction. The common principle is that AI is a tool and human judgment is the key. The question is how we move from AI to human judgment and recognise that there are things that cannot be replaced.

Another important principle is responsibility. Whatever you use AI for, you are responsible for the result. You cannot say, “Oh, this was the AI.” Responsibility remains human.

The questions Tanja and I are raising now need to be addressed if AI is going to become a responsible and reliable assistant. As part of the course, we hope to have participants assess different AI systems and LLMs and consider which they find most adequate and relevant for their work. We will also have facilitators who are currently dealing with conflicts, including people working in Afghanistan, who can present how they are using AI, what the benefits are and what the risks are.

I think our workshop is particularly useful for practitioners in the field as we know from our own experience what they need.