Artificial intelligence and technology (Getty images)
AFCEA TECHNET 2023 — In order to “catch up” to China in the artificial intelligence race, a senior Army acquisition official today charged industry with developing a bill of materials (BOM) as the service looks at increasing its adoption of third-party algorithms.
“The AI BOMs are critical for us and this is where we really want industry’s feedback because depending on what we say we want as part of the recipe for the AI BOM, y’all can come back and say, well that’s basically RIP…And this is where I think we want to have more dialogue,” Young Bang, principal deputy assistant secretary of the Army for acquisition, logistics and technology, said at the AFCEA TechNet Augusta conference.
AI BOMs, an idea first floated earlier this year by Bang, would be structured like software BOMs, or S-BOMs, used to secure the software supply chain. In essence, BOMs give a roadmap for where code inside key software applications came from — or as the National Telecommunications and Information Administration puts it, “a list of ingredients that make up software components,” which backers say reduces the cybersecurity risk.
The service wants to use industry’s algorithms to “catch up to China in this space,” but it will require looking at whatpotential risks or vulnerabilities those algorithms could create, he added. Bang emphasized that the idea isn’t to steal industry’s intellectual property or reverse engineer it, but to work together.
“This one, we’re still continuously talking to industry about and we want feedback because…we’re going down this path for the same reasons,” Bang said about AI BOMs. “We want to reduce our attack surface from an algorithmic standpoint.”
Bang has previously mentioned how reducing that risk is a concern, saying it’s a “hard challenge” that the Army needs industry’s help in. Another worry he’s floated before when it comes to the Army developing and integrating AI capabilities is an inexperienced workforce.
“Everyone’s interested in AI. We’re not there yet. Everyone’s interested in machine learning. Some people understand that,” Bang said May 25. “There’s not enough people in the Army… That’s why we need help from industry.”
AI BOMs will be one part of the Army’s digital transformation journey. During his keynote speech today, Bang briefly spoke about an approach called Traceability, Observability, Replaceability and Automated Consumption, or TORC, could help the service “liberate” its data.
Currently, the service’s systems are vertically integrated, starting with software at the top, then data, hardware and the network.
“Over time, this works for a system. But now we have multiple systems that are siloed…because it’s more efficient, or maybe it’s faster performance,” he said. “But when you look at it from an enterprise, it’s really not scalable…unintentionally it locks in that data into the systems and we literally have to perform miracles…to get into the system to pull out the relevant data. It’s tedious to get there. So what we’re talking about is how do we actually then decouple all of that?”
Now, the Army wants to separate those layers and use the TORC process to help it extract those layers to more easily get the data it needs, although Bang didn’t expand on the specifics of how that will work.
“What we’re trying to do is define the pieces in the middle of those extraction layers,” he added.