Publication

Artificial Intelligence, Non-Proliferation and Disarmament: A Compendium on the State of the Art

Cover of the publication with the text: EU Non-Proliferation and Disarmament Consortium, Promoting the European network of independent  non-proliferation and disarmament think tanks. No. 92. January 2025. ARTIFICIAL INTELLIGENCE,  NON-PROLIFERATION AND DISARMAMENT:  A COMPENDIUM ON THE STATE OF THE ART. thomas reinhold, elisabeth hoffberger-pippan, alexander blanchard,  marc-michael blum, filippa lentzos and alice saltini
Compendium by Thomas Reinhold, Elisabeth Hoffberger-Pippan, Alexander Blanchard, Marc-Michael Blum, Filippa Lentzos and Alice Saltini

This multiauthored compendium offers a state-of-the-art summary of the artificial intelligence (AI) issues facing non-proliferation and disarmament. It pulls together four topics—Artificial Intelligence in the Military Domain: Technical, Legal and Ethical Perspectives by Thomas Reinhold, Elisabeth Hoffberger-Pippan and Alexander Blanchard (section I); Artificial Intelligence and Chemical Weapons by Marc-Michael Blum (section II); Artificial Intelligence and Biological Weapons by Filippa Lentzos (section III); and Assessing the Implications of Integrating AI in Nuclear Decision-making Systems by Alice Saltini (section IV)—that, taken together, offer a concise overview of the proliferation- and disarmament-related challenges and opportunities that AI presents. Section I describes how military organizations increasingly use AI to enhance operational effectiveness in weapon systems, decision support and intelligence, and illuminates some of the critical technological, legal and ethical challenges posed by AI’s integration into military organizations. Section II examines AI’s impacts on chemical weapons, highlighting emerging risks from state and non-state actors, the need for regulation to prevent misuse and the importance of global collaboration to uphold norms against chemical warfare. Section III explores security concerns raised by the intersection of AI and biology with a specific focus on the risk that AI could facilitate the deliberate use of bacteria and viruses to inflict harm, emphasizing the need for a nuanced and evidencebased understanding of these risks. Finally, section IV examines AI integration into nuclear command, control and communications systems, noting its potential to enhance intelligence and situational awareness alongside significant risks of unreliability, cyber threats, and misaligned decision making, while calling for international dialogue and regulatory measures to avert catastrophic escalation. The texts compiled in this compendium were originally prepared as briefs in support of four ad hoc seminars on AI and arms control for the European Union and its member states.

Bibliographic record

Reinhold, Thomas; Hoffberger-Pippan, Elisabeth; Blanchard, Alexander; Blum, March-Michael; Lentzos, Filippa; Saltini, Alice. “Artificial Intelligence, Non-Proliferation and Disarmament: A Compendium on the State of the ArtNon-Proliferation and Disarmament Papers, 92. January 2024

Open Access publication

Project partners