Commentary Urges International Limits on Misuse of Autonomous AI Weapons

Summary

Matthew Rosenquist warns that autonomous AI weapons, including cyber variants, could be deliberately used against civilians, societies, and critical infrastructure, and urges international policymakers to establish boundaries before such capabilities scale.

Key points

• The author is more concerned about people directing AI systems to harm societies and political adversaries than about sentient AI acting independently.
• Autonomous weapons may have kinetic or cyber capabilities, according to the article.
• The author identifies attacks on non-combatants, humanitarian targets, and supporting critical infrastructure as severe risks.
• The article calls for international policymakers to agree on acceptable boundaries and responses to the deliberate use of such weapons against protected targets.
• It warns that authoritarian or desperate governments may pursue these capabilities at scale.