Research Summary: Autonomous Cyber Weapons as AI-Enabled Digital Harms: A Platform Accountability Framework

A research paper proposes a platform accountability framework to address the governance gap regarding state-deployed autonomous cyber weapons targeting civilian infrastructure, which are omitted from commercial regulations like the EU AI Act and Digital Services Act. The framework introduces a digital-harm taxonomy, a Meaningful Human Control model tied to International Humanitarian Law thresholds, and draft treaty language based on the Articles on Responsibility of States for Internationally Wrongful Acts.

Why it matters

This framework matters because critical civilian infrastructure is increasingly vulnerable to sophisticated state-sponsored autonomous cyber attacks that fall outside current regulatory regimes. Establishing formalized, legally grounded mechanisms for state accountability is essential for defining international norms and deterrence in algorithmic warfare.

Key insights

• Existing platform and AI regulations, such as the EU AI Act and UK Online Safety Act, leave state-deployed autonomous cyber weapons against civilian assets unregulated.
• The authors develop a digital-harm taxonomy crossing six autonomous cyber weapon operational modes against five distinct harm categories.
• A formal Meaningful Human Control model is introduced, featuring a closed-form International Humanitarian Law compliance threshold and a proved weight-sensitivity bound.
• The governance architecture leverages the ARSIWA state-responsibility doctrine for international accountability rather than individual criminal liability under the Rome Statute.
• The framework was evaluated against eleven documented state-attributed operations and a hypothetical case.

Source

arXiv — Computers and Society — https://arxiv.org/abs/2610.32238