Recap And Reflections From The Advantage DoD Defense Data & AI Symposium
Representing Parsons At Advantage DoD 2024
I had the honor of representing Parsons Corporation at the inaugural Advantage DoD Symposium, hosted at the Washington Hilton last week. The invitation came as the result of a solution our AI Bench, under the SPARC organization within D&I, devised through the lens of risk management for the comprehensive evaluation of Large Language Models (LLMs), systematically gauging their maturity to handle DoD-specific use cases, and mitigations for immature and imperfect models. Dr. William Streilein, CDAO Chief Technology Officer, commended Parsons on our comprehensive approach to the design of an LLM Maturity Model and asked that we share our concept with symposium attendees. Positive reception by attendees representing both the Defense industry, private enterprises, and academia and earned Parsons an invitation to participate in a consultative capacity in a DoD-led knowledge consortium to follow, cementing Parsons’ future a thought leader in AI.
The Current State Of AI In The DoD
The Office of the CDAO set the stage with an overview of the DoD’s current exploratory yet experimental phase in leveraging AI to bolster operational efficiencies. Although a large focus of the conference was on LLMs, the topics discussed encompassed the gamut of topics in AI, emphasizing their commitment to harnessing cutting-edge technologies for a competitive edge and enduring decision advantage. The CDAO emphasized evaluating industry advancements for DoD-specific applications and fostering collaboration for effective technology adoption while the DoD builds up a fighting force that contains intelligent practitioners of adopted AI technologies.
What The DoD Needs
The CDAO’s technology goals utilize AI to enhance warfighting and support tasks, with a focus on using industry tools and techniques to increase efficiency rather than pouring resources into building proprietary models. Using AI to free up human resource capacity and unify disparate and siloed knowledge systems are among some of the many needs discussed by CDAO at the conference. The CDAO’s 18-month generative AI task force, Task Force Lima, was established in August 2023 with the aim of developing foundational strategies for the responsible integration of generative technologies. CAPT Xavier Lugo, TF Lima Mission Commander, warned that the DoD is not interested in more chatbot use cases but rather invested in harnessing LLMs’ knowledge and reasoning abilities to maintain a competitive warfighting edge. The DoD needs methods for evaluating the potential risks and consequences of augmenting existing workstreams with AI technologies, strategies for mitigating those risks, and roadmaps for department-wide adoption and synchronization of the policies for those technologies.
Where AI In The DoD Is Headed
It is abundantly clear that the DoD needs trusted leaders in industry to support the needs of an AI-powered, interconnected warfighting support system. The Deputy Secretary of Defense’s 2023 DoD Data, Analytics, and AI Adoption Strategy prioritizes speed of delivery, continuous learning, and responsible development in its agile approach to this technological transformation. Therefore, a concerted effort by military leaders, servicemembers, and forward-thinking industry partners like Parsons will be necessary to achieve the goals of the DoD’s AI adoption strategy. A Parsons-DoD partnership is an optimal application of our experience in applying AI to the most critical aspects of military and intelligence operations, keeping our fighting forces at the cutting edge of technological innovation and ultimately enabling DoD leaders to make rapid decisions by expertly leveraging high-quality data as part of a continuous development cycle.