Contested Logistics: How AI Can Help Protect and Defend
Over the past two decades, warfare has changed. Through observation and AI-enabled analysis, our adversaries know where we are and what we’re trying to do. We know the same about them. So, we also know that one of their primary targets in the US is our critical infrastructure, including logistics.
On Thursday, March 20, a panel of high-level intelligence, military and technology leaders convened in McLean, VA to discuss “Dynamic Predictive Analytics for Contested Logistics.” The session was hosted by The Potomac Officer’s Club and it was moderated by SAIC. Primer’s Skip Farmer, Senior Director, Field Engineering, was one of the panelists.
The discussion centered around contested logistics—situations where the transport of goods and services intended for the warfighter are hindered or disrupted by an adversary, impacting the ability to sustain military operations. Whether this manifests in physical attacks, cyberattacks, interdiction or information warfare, data forms the foundation.
Using AI in contested logistics.
One of the main discussions centered around new national defense strategies and the importance of leveraging technology to address threats. Rather than transforming contact progressively over the next several years, the timeline is now focused on doing that in the next zero to three years. It requires a rethinking of programs, leveraging of technologies and shortening ATO cycles so tools get into the hands of warfighters faster.
One of the greatest advantages of AI to defense is its ability to pore through millions of documents in real time and deliver a summary and analysis supported by data. In discussing the challenges and opportunities of AI in contested logics, the experts cited:
- Context: By analyzing contested logistics, we gain a clearer understanding of our adversaries’ movements and technological capabilities. This understanding informs our position in the AI landscape and highlights areas of risk or opportunities for improvements.
- Overwhelm. Analysts are often beset with cognitive overload from all the data they have sort through. Meanwhile more and more data is rolling in. AI helps analysts get insights and summarize information faster and closer to real time.
- Surprise. We plan for certain things in conflict. But by analyzing publicly available information and internet chatter, we can unearth elements we are not planning for—and elements our adversaries are not planning for. AI can reduce the element of surprise so we can plan both offensive and defensive strategies with more confidence and strength. It can also help account for the use of deception in the process.
- Cost. Military leaders understand the requirements on the table in terms of technology. But in an era of shrinking budgets, leaders need quick wins to learn from those experiences, as well as to justify costs.
- Partnership. The government can’t do this alone. So, there needs to be more partnerships with industry to develop and refine bespoke solutions.
- Trust. Trust is the five-letter word that underpins the analyst’s entire existence. One simple error can have huge implications. Recent developments in RAG-Verification are improving trust in the technology.
- AI at the Edge. Panelists stressed a need to bring AI to the edge—to make it feasible for the warfighter in the field. This is currently difficult because of AI’s compute and bandwidth needs, which are frequently limited in the field. There is a need for coordination between HQ and the edge to get analysis to warfighters faster.
Along the way, panelists mentioned some of the capabilities on their AI wish lists, such as using it for predictive maintenance, risk management and sensor placement optimization. There was also a discussion of the value of failure—specifically failure to learn lessons and inform future analysis.
Modernizing data and AI in the DoD.
Near the end of the discussion, Project Convergence came up. Project Convergence is the Army’s mission of experimenting with new technologies, then integrating the winners across all domains, rather than in silos, across all domestic forces and some allied countries. This joint command and control approach delivers a unified defense and offence in times of battle, with unified intelligence and tactics.
One senior military leader suggested that AI companies should not be afraid to have frank conversations with defense prospects. The military has promoted Project Convergence and it’s not a sensitive topic. However, they shouldn’t approach the pitch guns blazing, but from a more broad and neutral point, warming up to what you can do for them. Keep the military’s priorities in mind and come bearing ideas.
The session was attended by hundreds of government and industry leaders and the hour was filled with insights and smart questions. The conclusion is that the DoD sees AI as integral for their plans of contact transformation, strategic surprise, alleviating data overwhelm and addressing contested communications. However, more work is needed in both the technological and adoption aspects. Strong partnerships between defense and industry are required to shape capabilities to needs.
Primer is always open to discussions regarding current capabilities and future development in support of US military. Contact us and let’s get the innovation rolling.
