Can AI Make Conflicts Worse? An Alignment Failure in LLM Deployment Across Conflict Contexts

TL;DR

This paper evaluates how current large language models perform in conflict-sensitive scenarios, revealing significant risks of misinformation and division, and introduces a new framework for assessing alignment in such contexts.

Contribution

It presents the first evaluation framework for assessing LLM behavior in conflict scenarios and highlights the importance of model choice for safety.

Findings

Failure rates range from 6% to 47% across models.

Five configurations failed 80-100% in sensitive cases.

Model choice significantly impacts safety in conflict contexts.

Abstract

AI models are already deployed in societies affected by armed conflict, and journalists, humanitarian workers, governments and ordinary citizens rely on them for information or for their work processes. No established practice exists for checking whether their outputs can make those conflicts worse. We tested nine model configurations from four providers (OpenAI, Anthropic, DeepSeek, xAI) on 90 multi-turn scenarios designed to surface misaligned behaviour in conflict contexts: false equivalence between documented atrocities, denial of genocide, and failure to recognise ethnic slurs, among others. When such outputs feed into journalism, humanitarian reporting, or public debate, they can deepen divisions in fragile societies. Failure rates span 6\% to 47\% between the best and worst performing models, which makes model choice a safety question in its own right and when users pushed for ``balance'' in cases where international courts have already assigned responsibility, five of nine configurations failed 80 to 100 percent of the time. We release the first evaluation framework for this domain and propose adding it to alignment evaluation portfolios.