Building trust in data and AI for conflict prevention: Three lessons from the Data for Peace Conference 2026

As artificial intelligence becomes increasingly integrated into conflict prevention, trust is emerging as the defining challenge. Trust in data. Trust in analysis. Trust in institutions. The Secretary-General's New Agenda for Peace reminds us that trust is the cornerstone of collective security. As AI enters peace and security practice, the question is no longer whether we should use it, but how it can be used in ways that strengthen political judgement, accountability and decision-making. 

These questions were at the centre of a panel co-organized by UNSSC and the DPPA Innovation Cell during the Data for Peace Conference 2026. Practitioners from across the UN system and civil society, including DPPA-DPO Information Management Unit, UNOCC, UNDP and BuildUp, reflected on how AI is reshaping the use of data in peace and security work. Rather than debating whether AI is inherently “good” or “bad”, the discussion focused on where trust is created or lost, where risks accumulate and which parts of the data-to-decision chain require stronger protection as AI becomes more prevalent. Three lessons emerged: 

1. Lesson 1 : Trust in data - Trustworthy AI begins with trustworthy data  

AI systems are as reliable as the data and assumptions behind them. In conflict settings, incomplete, biased or manipulated information can have significant operational and political consequences, including distorting early warning, misidentifying drivers of violence, and undermine confidence in preventive engagement. While AI can support data collection and quality assurance, it cannot compensate for poor-quality data and may amplify existing weaknesses. This makes data provenance, contextual validation and quality control essential to conflict analysis. Even manipulated or misleading information can retain analytical value when assessed as a signal of actor intent, tactics, narratives or emerging threat vectors. Data should therefore be treated not only as evidence, but also as a space where political contestation unfolds.  

2. Lesson 2 : Trust in human oversight : Keeping political judgement at the centre  

Human analysis is essential throughout the data cycle to ensure AI-supported insights are accurate, relevant and actionable. A conflict analyst may identify new trends or emerging tensions that are not yet visible in historical datasets but are understood through local relationships, political judgement and contextual knowledge. Rather than replacing human judgement, AI should augment the expertise of analysts who understand the incentives, grievances, histories and social dynamics shaping conflict. Maintaining meaningful human oversight also preserves accountability: decisions affecting peace and security must remain explainable and politically grounded, not delegated to an opaque analytical process.   

3. Lesson 3 : Trust in institutions : Bridging global frameworks and field practice 

Trust in AI for conflict prevention also depends on responsible governance frameworks that address misinformation, bias and poor-quality data. Without such safeguards, flawed inputs can lead to less inclusive analysis, mistimed engagement and misjudged risks. Under the Global Digital Compact, adopted as part of the Pact for the Future at the Summit of the Future in September 2024, Member States agreed to establish two new mechanisms on AI: an Independent International Scientific Panel on AI and a Global Dialogue on AI Governance. These mechanisms were formally established by the General Assembly in August 2025, to provide independent scientific assessments and a multistakeholder platform for international cooperation on AI governance, including to help close digital divides and promote equitable access to the benefits of AI. Turning global commitments into everyday practice requires sustained investment in capacity-building. Institutions such as UNSSC, working alongside departments and entities across the UN system, including DPPA, UNOCC, and UNDP, can help translate global frameworks into practical guidance while ensuring field-level experience informs global discussions.  

A consistent message emerged from the discussion: trust is central to the responsible use of data and AI in peace and security work. The future of AI in conflict prevention will depend not only on technological advances, but on our collective ability to build trusted data ecosystems, preserve human judgement and strengthen responsible institutional practices.