When Mediators Need Machines (and Vice Versa): Towards a Research Agenda on Hybrid Peacemaking Intelligence

Abstract

The growing use of artificial intelligence (AI) in peace processes challenges the ways in which mediators conventionally generate and use knowledge to facilitate a convergence of conflict party positions. Mediators are increasingly unable to maintain their role as information brokers without the support of automated information gathering and analysis systems. However, the effectiveness of such AI-tools is likewise hampered by their limited autonomy across peacemaking contexts, by the paradigmatic nature of knowledge that is used to design AI-systems, and the influence of subjective factors that are difficult to measure. These challenges may be overcome by hybrid human-machine systems in support of peacemaking. This article formulates a research agenda for hybrid peacemaking intelligence that is concerned with the increasing interdependence of humans and machines. It suggests studying how agency is distributed in mediator-machine networks, the hermeneutics of machine-supported participatory approaches, and how hybrid arrangements change power relations in peacemaking.

Supporting peace negotiations in the Yemen war through machine learning

“…A deep understanding of a conflict situation, the key conflict areas and how they may be interrelated, its actors and their interests, the relationship between different actors, and their potential openness to finding alternative solutions is key to devising an effective dialogue strategy (Amon et al, 2018). Machine learning tools may be particularly useful in complex and lengthy mediation processes in which viewpoints may otherwise calcify and prevent consensus finding by helping to challenge stereotypes and prejudices that may have built up over time (see Hirblinger, 2022, p. 20).…”

“…Complex and sensitive mediation processes remain a human-centered trade for which machine learning tools can only offer a degree of analytical support. At present, ensuring “meaningful human control” (Höne, 2019, p. 11) over what Hirblinger (2022) refers to as “hybrid intelligence peace-making systems” is best achieved when mediators are actively involved as collaborators throughout the development process. Continuing to follow a participatory, cocreation methodology will also help the mediation team to develop further their understanding of the kind of problems that machine learning tools can provide assistance with.…”

“…It was clear from the feedback from the mediation team that their trust in the tools developed had its foundations in the project's participatory, cocreation methodology, which enabled them to query how the tools work, seek clarification about the meaning of the results they generated and thereby retain the necessary oversight of the mediation process (Hirblinger, 2022). As part of ongoing collaboration with our collaborators, we will focus on how to create and present explanations of the analyses produced by the tools that will satisfy these requirements.…”

Digital Peacebuilding: A Framework for Critical–Reflexive Engagement

“…Finally, suggesting that digital peacebuilding creates practices, forms, or rationalities of government that are "artificial" or "post-human" would mean to isolate digital peacebuilding from its sociotechnological environment. However, this would create an artificial human-machine distinction, as digital technologies, including AI, continue to be developed, used, and controlled by humans ( Hirblinger 2022 ).…”

Leveraging Digital Methods in the Quest for Peaceful Futures: The Interplay of Sincere and Subjunctive Technology Affordances in Peace Mediation

“…Data-driven methods often have a 'scientific' appeal that all too easily raises the expectation that they will enable more objective and accurate analysis and decision-making. However, 'chasing the future' requires considerable human involvement, given the limited accuracy of even the most advanced AI-supported conflict prediction models (Hirblinger, 2022). Importantly, conflict parties and stakeholders included in data gathering and analysis exercises will usually hold strongly subjective views on the conflict and may associate data and information with hurtful grievances and conflicting stances, which makes their utility for peace mediation by no means straightforward.…”