IMI-Aktuell 2024/574
KI: Wettrüsten
(05.09.2024)
Die Untersuchung „Escalation Risks from Language Models in Military and Diplomatic Decision-Making“ beschäftigte sich mit dem Verhalten von KI-Systemen in Kriegssimulationen und warnt auf Basis der Ergebnisse vor deren Einbeziehung in die Entscheidungsfindung: „Governments are increasingly considering integrating autonomous AI agents in high-stakes military and foreign-policy decision-making, especially with the emergence of advanced generative AI models like GPT-4. Our work aims to scrutinize the behavior of multiple AI agents in simulated wargames […]. We observe that models tend to develop arms-race dynamics, leading to greater conflict, and in rare cases, even to the deployment of nuclear weapons.“ (jw)