Abstract for: Nat-Sec MAISC Research Agenda: Challenges & way ahead for M&S, AI, on topics of systemic complexity in national security
Despite GenAI's rapid growth in many areas, deep problems of systemic complexity and the compute-power problem present limits to what the scaling hypothesis may achieve in GenAI only. Integrating M&S with AI is necessary to advance capabilities for systemic complexity topics such as national security. Three DoD-sponsored expert-attended workshops in the US & UK, as well as individual interviews of over 90 M&S, AI, and systemic complexity experts. A common limitation is that views represent those invited and able to attend workshops, conferences, or interviews. Findings, therefore, should not be taken as a statistically valid representation of any field or perspective but as notional and illustrative of the needs of Nat-Sec MAISC. Problem-oriented research agenda that can shape future efforts to solve the four deep problems of systemic complexity. Organized across nine problem topic areas and over 50 sub-topic areas. Additionally, eight simulation-in-the-loop recommendations and an organizing construct for the way ahead is offered. The Nat-Sec MAISC Agenda is larger than any single firm, university, defense contractor and possibly even nation-state actor. It may take multilateral cooperation, such as between Five-Eye countries, and an ecosystem of technology firms, academia, and government agencies organized in an effort similar to a Manhattan Project or moonshot effort.