Automated LLM-based Intelligence Compilation and Evaluation (ALICE)

Military mission success is directly impacted by the quality and recency of the intelligence supporting the decision-making process. Currently, the Marine Corps Planning Process (MCPP) approach to Intelligence Preparation of the Battlespace (IPB) resulting in the generation and vetting of information for a daily intelligence report is time-consuming, manpower intensive, and requires significant cognitive load. The IPB process requires information to be collected and correlated from a wide variety of locations and modalities to generate the content required for the IPBĺs wide range of topics including Defining the Operational Environment, Effects on Operations, Friendly Capabilities, Evaluation of the Adversary, Adversarial Course of Action Analysis, Information Requirements (IR/PIR/SIR) development, and Effective Exploitation (e.g. Cueing, Redundancy, Mix, Integration, and/or Coordination). A system that was capable of automatically finding the best sources, extracting key information, correlating the information across data sources and modalities, and providing rapid intuitive access to the derived knowledge in a way that enables the end consumer to understand the reliability of the information would be of tremendous benefit to the military. Therefore, the Machina Cognita Technologies (MCT) Team including Covan Group and Serco propose to develop the Automated LLM-based Intelligence Compilation and Evaluation (ALICE) system. The ALICE system will ingest, parse, tag, analyze, and correlate the available data sources and make them easily accessible via a REST API.