Papers of BAS. Humanities and Social Sciences

Vol. 8, 2021, No. 1

Artificial intelligence and its possible use

in international nuclear security law

Anguel Anastassov

Abstract. The paper evaluates the possible implications of artificial intelligence (AI) in some key international nuclear treaties. AI could play a substantive role in the detection of cyberattacks on nuclear systems. However, AI is still an emerging technology that does not allow yet removing the human factor from the decision-making processes in the modern nuclear security realities. States have not reached yet a common understanding about the degree to which current international law would be able to regulate AI developments. A conclusion is made that the prospective first steps in using AI could be to adopt mutually agreed upon arrangements at the level of producers, consumers, and the general public, as well as to develop methodologies to certify AI products. The obvious challenges are related to AI’s dependence on a representative, sufficiently large set of data with good quality.

Keywords: artificial intelligence, nuclear stability, nuclear security law, regulatory gap