---
title: How malicious AI swarms could destabilize states
author: Martin Ebner
date: 2026-11-12
url: https://postgenerative.at/indications/2026-11-12/how-malicious-ai-swarms-could-destabilize-states
lang: en
type: Indication
summary: A preprint examines possible destabilization pathways through Mali and Ethiopia.
tags:
  - Politics
  - Media & Public Sphere
  - Security
sources:
  - title: "Schroeder et al. (2026), Preprint: Automated Disinformation and Malicious AI Swarms: Risks for Democracy and Development in Africa"
    url: https://arxiv.org/abs/2610.11930
  - title: Volltext (PDF)
    url: https://arxiv.org/pdf/2610.11930
license: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)
alternate:
  de: https://postgenerativ.at/indizien/2026-11-12/wie-boesartige-ki-schwaerme-staaten-destabilisieren-koennten
---

# How malicious AI swarms could destabilize states

Coordinated, adaptive AI agents could fabricate consensus, inflame identity conflicts and undermine institutional legitimacy. In Mali, they could amplify competing conflict narratives; in Ethiopia, augment existing state propaganda networks. Linguistic data gaps also weaken defences. [1]

**The evidence boundary:** Large-scale campaigns by fully autonomous swarms remain unproven. This forward-looking risk analysis identifies human-directed AI campaigns as a particular near-term threat. It recommends platform accountability, independent local media and regional cooperation while protecting political freedoms. [1]