By Kaleab Azeze
Ethiopia’s transitional justice process is unfolding in a deeply polarised environment, where competing narratives about past conflicts continue to shape public trust, political identity, and even disagreement over basic facts. In such a context, transitional justice risks being seen not as a shared national effort toward accountability and reconciliation, but as another arena of political struggle.
This is not unique to Ethiopia. In highly divided societies, polarisation can operate as a “hyper-problem” that weakens institutions, deepens mistrust, and increases the risk that justice processes become politicised rather than unifying (Institute for Integrated Transitions, 2021).
At the same time, transitional justice mechanisms—truth commissions, prosecutions, reparations, and institutional reforms—are highly sensitive to perception. Even well-intentioned processes can unintentionally deepen divisions if they are seen as biased or exclusionary.
This is why Ethiopia should consider integrating artificial intelligence (AI) prediction tools into its transitional justice process.
AI systems are already being explored in governance and conflict prevention to anticipate risks and support decision-making in fragile contexts (United Nations University, 2024). In transitional justice, similar tools could help policymakers anticipate how different communities are likely to react to proposed measures before they are implemented.
For example, AI could help identify whether a truth-seeking process will be perceived as inclusive or partisan, whether accountability measures risk being framed as selective justice, or whether reparations may unintentionally reinforce exclusion.
This is not about replacing human judgment. AI should function as a decision-support and early-warning tool, helping institutions understand the likely social consequences of their choices before they escalate into conflict (Government Information Quarterly, 2025).
However, safeguards are essential. AI systems are not neutral and can reproduce bias or undermine legitimacy if poorly designed. They must never determine legal responsibility or replace democratic decision-making (United Nations, 2023).
The question is therefore not whether technology belongs in transitional justice, but whether Ethiopia can afford to design such a sensitive national process without tools that help anticipate how society will respond.
Used responsibly, AI prediction tools could help Ethiopia build a more adaptive and conflict-sensitive transitional justice process—one that strengthens legitimacy and reduces the risk of deepening division.
References
Institute for Integrated Transitions (IFIT). (2021). Polarisation: The ‘Hyper-Problem’ Transitional Justice Can No Longer Ignore.
United Nations University. (2024). Predictive Technologies for Conflict Prevention: Practical and Policy Considerations. United Nations University (UNU).
Government Information Quarterly. (2025). AI in Public Sector Decision-Making: Opportunities and Governance Challenges.
United Nations. (2023). Guidance Note of the Secretary-General on Transitional Justice. United Nations.
