J. Babiec, H.Snyder
Via Science, Inc., Massachusetts, United States
Keywords: Generative AI, AI/ML, Data Integration, Data Visualization, Data Security
Traditional dashboard-based workflows, the backbone of DoD data-driven decision-making, are inflexible and increasingly insufficient to match the operational tempo and uncertain requirements of future multi-domain operations involving apex adversaries. DoD’s data challenge will increase exponentially with the proliferation of smart sensors and equipment for an expanding range of intelligent systems and systems of systems. VIA's SLAM-AI application (Share Less Ask More) complements existing dashboards to accelerate the speed, security, and quality of DoD decision-making. SLAM-AI leverages GenAI to convert human-language prompts into custom visualizations on demand. Unlike all other GenAI tools, SLAM-AI never shares data with a GenAI model so the model cannot learn from or leak sensitive information. Data-centric permissions and a zero trust architecture automatically manage each user’s authorized activities, while query auditability detects and deters unauthorized activity. Post-quantum encryption provides extreme cybersecurity resilience. SLAM-AI has cATO (IL4; IL5 pending). It deploys on US military cloud infrastructure, commercial clouds, and in on-prem or air-gapped configurations. It is robust in DDIL environments. SAF/IEE and US military installation managers use SLAM-AI to accelerate data-driven decision-making for critical missions, validating the capability’s reliable performance, rapid overlay deployment into existing enterprise environments, and overall security.