AI Overview

It’s always been about the data and the analytic component on top of that — making sure you are a critical data organization, have data management, and data security and auditability.

Aki Jain

Process Changes with AI

With an exercise like Scarlet Dragon, it’s how do we make 1,000 decisions in an hour? It's ultimately that end-user goal.

Aki Jain

Agencies have only begun to realize the power and potential of artificial intelligence. Over the past few years, the Defense Department specifically has invested a great deal of time, money and effort into laying the foundation for better use of its data through the adoption of AI tools and capabilities.

Starting in 2017, DoD saw the handwriting on the wall as near-peer competitors like China and Russia accelerated their efforts to apply AI to weapons and systems.

The Congressional Research Office said back in November 2020 that DoD must work through distinct challenges to further take advantage of AI’s clear potential benefits to military operations.

Among the questions CRS posed were three that focus on unique AI use cases at the department:

• What types of military AI applications are possible, and what limits, if any, should be imposed?
• What unique advantages and vulnerabilities come with employing AI for defense?
• How will AI change warfare, and what influence will it have on the military balance with U.S. competitors?

Artificial intelligence as a means to an end

It’s important to keep in mind when talking about AI, or really any emerging technology, that it is ultimately a means to an end, said Aki Jain, president of U.S. government at Palantir.

“How do I ultimately enable my warfighter, my analyst, my operator to leverage that data in order to have some kind of decision advantage? That’s the first thing. That’s really the use case that we’re seeing for AI,” Jain said during a Federal News Network  discussion on “Technology and the Great Power Competition: AI complexity and competition.”

“The second one is really thinking of AI as a teammate, just like you have other folks that you work with on a team to achieve some kind of outcome,” he continued. “How do I ultimately enable AI to be more seamless, to be a trusted teammate and solve a problem with me? Sometimes that’s dealing with the deluge of data, sometimes it’s just helping me look at new perspectives with that data that might suggest a clear course of action.”

Using AI to dive deep into DoD work streams

From Jain’s perspective, those are the two high-level takes on Defense use cases. Then, there are hundreds of work streams that agencies within DoD could pursue for each by applying AI, he said.

Over the past several years, the understanding and application of AI across DoD mission areas has become clear.

Whether it’s the basic use of AI in robotics process automation (RPA) for financial management services or predictive analytics to help meet airplane logistical needs, DoD is using these technologies to help automate, enhance and better serve service members and civilians.

The many different flavors of AI offer potential improvements, but DoD, or any organization for that matter, can’t just apply the technology across the board, Jain said. To better understand where AI can have the biggest impact, agencies should start with their unique data.

“It’s always been about the data and the analytic component on top of that — making sure you are a critical data organization, have data management, and data security and auditability. That has always been a key part of the AI equation,” Jain said.

The convergence of cloud computing and pervasive graphics processing units (GPU) access has given organizations a set of problems that are ideal for applying AI and machine learning (ML) techniques, he said. “At the core of what we really believe is that whether it’s a human, another machine or an algorithm, what you’re trying to ultimately do is enable some kind of collaboration across all those parties to enable them to make a better combined decision to drive some kind of outcome.”

Real-world AI testbeds

DoD has begun to test out that type of AI- and ML-led collaboration through exercises like Scarlet Dragon and Project Convergence.

These and other efforts are creating a shift in how DoD thinks about and applies AI and ML to solve problems, Jain said.

“With an exercise like Scarlet Dragon, it’s how do we make 1,000 decisions in an hour? It’s ultimately that end-user goal that they’re trying to achieve,” he said. “When I think of things like Project Convergence, they are really a little bit more kind of forward-leaning. They’re trying to think about where do we need to be three to five years from now to have that decision advantage?”

In many cases, the department is focused on the deterrence advantage, he said. “There are aspects that I can’t speak to you about, but the things that I am aware of are really cool, and they’re impressive.”

A key piece to this broader understanding of what AI can do will evolve through the partnership with industry, Jain said, adding that innovation going on across the private sector can help DoD progress toward its end goal faster.

“It’s not about Silicon Valley versus Washington or anything like that. There’s lots of great companies and extremely, extremely talented engineers that want to support and enable this mission all over the country. We should be taking advantage of them and creating opportunities for them,” he said.

DoD is working to embrace AI as a mechanism for deterrence, for instance in information dominance. “You need all-domain command and control in a scenario where you are in conflict, but the deterrent comes from building those advanced capabilities.”

Listen to the full show:

Copyright © 2024 Federal News Network. All rights reserved. This website is not intended for users located within the European Economic Area.