On 22–23 September, SIPRI organized a workshop in Stockholm to discuss the impact of military artificial intelligence (AI) on nuclear escalation.
The event gathered diplomats and experts on nuclear, military AI and arms control issues—including from seven of the nine nuclear-armed states—for two days of in-depth discussions. These sessions focused on whether the military use of AI creates new escalation risks or exacerbates existing risks. The workshop was structured around three scenarios designed to explore the impact of AI on nuclear escalation. The purpose was to create a better and more nuanced understanding of the risk landscape, with the goal of developing specific risk-reduction measures in the age of military AI.
Sessions were led by researchers from the SIPRI Weapons of Mass Destruction Programme and the Governance of AI Programme. The workshop was funded by the Carnegie Corporation of New York, and the information gathered during the workshop will feed into a forthcoming SIPRI publication on the topic.
Click here to read a recent SIPRI Insights Paper on the topic.