Author: Rose de Fremery
Date published: October 29, 2024
Military artificial intelligence (AI) is already transforming battlefield operations, providing near real-time insight into on-the-ground conditions so that commanders can make better decisions. Although this is a powerful example, the use cases for AI in defense could be far more expansive and profound than what we've seen so far. While AI is still nascent in terms of its proliferation, there is still much debate on compliance and how to monitor and deploy it responsibly.
This article will be exploring how the military is already using AI to enhance productivity and drive digital transformation, and how it can further leverage AI for even more military innovative use cases in the future.
The United States military has a long track record of driving technological innovation. In the 1960s, for example, the Defense Advanced Research Projects Agency (DARPA) created a computer network known as ARPANET that served as a conceptual framework for the internet.1 This kind of innovation has continued over the years, as DARPA remains the place to take on exploring new technology in context of the unique sensitivities of military requirements.
Using AI in military operations can support several advantages; fast, precise and resilient kill chains is one example.
"That specifically refers to a threat, whether it's in the battlefield or within your network and a cybersecurity operation," explained Dominic Bonaduce, senior manager for product strategy, technology and development at Verizon."If there's an attack, where's that attack happening? And how do I prevent that attack from spreading and causing more damage?"
By leveraging the integration of visual data from multiple sources along with archival data showing change over time AI empowers the ability to pinpoint what is happening, plus where it happened. This can help ensure that it doesn't spread to other vectors.
In November 2023, the DoD laid out its vision for AI. The strategy embraces agile military use of AI, placing a special focus on delivery speed and adoption at scale to facilitate the following desired outcomes:
"It shouldn't be a surprise that the Department of Defense (DoD) and the military have had a relationship with AI or artificial intelligence," said Bonaduce.
Experts are already using AI to analyze satellite images and drone video feeds faster than before. This visual intelligence can be particularly valuable on the battlefield, helping military forces understand the current state of play. This translates directly into the capability to react more quickly in critical fast moving situations in the field.
"If you have visual intelligence from satellite imagery over the course of five or 10 years or even 5 to 10 days, and you're looking to see what's changed in those images, AI could help you," said Bonaduce. "Whether you're looking for construction projects, land mass changes or weather pattern changes, you're going to start utilizing AI to look at those visuals, find where the changes happened, geolocate them and then identify what the changes may be." This visual intelligence can be particularly valuable on the battlefield, helping military forces understand the current state of play and provide insight that allows for faster and more accurate decision-making.
With the advent of generative AI, intelligence analysts can not only see the visual analysis that these AI models provide, but also are able to explore alternatives by asking questions about it.
"You might ask the AI, 'What changes have been observed over X number of years in the region of this photo?'," Bonaduce explained. This kind of tool could provide considerable advantages for the analysts who assess this form of intelligence, allowing for production of high quality analysis involving larger data sets and less time. If correctly applied, this military use of AI could accelerate analysis delivery without compromising quality. That said, the workforce needs proper education and training to ensure successful AI use in the military—and that in itself requires a high degree of forethought, planning and dedication.
Any time you implement some form of digital transformation, training and education are crucial to success. This is especially true for artificial intelligence. While it might appear at first glance that AI is easy to use, the reality is more complex. Leaders within the DoD who are charged with AI adoption and military innovation must focus on making sure that AI is leveraged correctly, not simply used as an accelerant.
"It all begins with effective digital transformation and workforce planning," Bonaduce said.
Creating a security-first, digital transformation strategy that utilizes modern infrastructure, software-defined networking and cloud-based controls can make it easier for organizations of all sizes to deploy new applications and services, including artificial intelligence. Take this assessment to learn what stage your organization is in.
The military has been developing internal expertise and proficiency with AI to allow the development of forward-thinking use cases to help further military innovation. According to Bonaduce, two especially promising future use cases for AI in defense focus on software development and IT support.
It’s no secret that AI coding tools are being used to enhance productivity, coding efficiency, and optimize coding workflows in software development. AI coding tools can create task automation capabilities, edit code, and provide contextual suggestions and this extends to military AI uses.
The DoD is using AI to accelerate software development. "Going back to that example of analyzing an image for intelligence purposes, that's software that has to be written," Bonaduce explained. "AI is going to be a framework and model that is incorporated into software to give it that capability. But that software still has to be created, and so AI can be used as part of the development process. You can use generative AI to give prompts for the types of software you're trying to build, so that your lines of code are written for you. You can use AI to then validate and debug that code and speed up the development time for these applications and software that are being employed and deployed by the military." AI pushes the limits of military innovation and enhanced operations.
Large Action Models (LAMs) are a type of artificial intelligence that are designed to understand user instructions and can perform tasks by interacting with different interfaces and systems. LAMs are an expansion and an improvement upon large language models (LLMs) which perform text generation, translation and can create summaries.
AI can help the military with daily IT support and troubleshooting and other instruction based operations. Just like in the private sector, DoD personnel run into technical glitches. A computer may stop working properly, a software update may fail to install or someone's phone operating system may freeze. In many cases, DoD personnel would need help from an IT specialist. However, IT specialist resources are often in short supply, so they might be prioritizing more urgent issues. LLMs, LAMs and other AI-powered tools can help reduce the stress on IT and networking operations.
"Being able to fully articulate and understand your issue is halfway to solving it," said Bonaduce. "If you can use generative AI to suss out and troubleshoot some of the problems that you may be experiencing, you'll free up technicians to solve other problems. Conversely, generative AI may also be used to help articulate what the issue is when that issue needs to be escalated."
Ultimately, AI has the potential to play a valuable supporting role in the military's day-to-day operations, as well as its mission planning and operational challenges.
Military use of AI, both now and in the future, may rely on consistent access to high quality data—which might be coming from anywhere at any time, including from bandwidth-intensive sources like drone video feeds, for example. Ultra-fast, secure and reliable network connectivity can be critical to the success of these AI use cases. 5G connectivity— such as the kind currently being used as an AI incubator for the DoD—can meet the stringent requirements needed for certain uses of AI use in the military.
Some military organizations may still need to look at their network infrastructure to make sure they have the network to support the data gathering capabilities that AI applications need before they can embrace AI technology.
"The reality is, a lot of infrastructure across the DoD includes a lot of copper that needs to be replaced with fiber," said Bonaduce. "There are a lot of TDM-based circuits that need to be switched to IP. There's a road to modernization that needs to be traveled first before you see the full proliferation of AI across an office setting or a military installation. Things like 5G are going to be critical to those types of capabilities."
Whether it's enabling fast and resilient kill chains or helping intelligence analysts interpret visual images faster, AI is already making an impact on military operations. In the future, the value of AI may extend to even more innovative use cases, such as internal software development and IT support. A robust forward-thinking network infrastructure is needed to maximize AI's potential for military use today, tomorrow and beyond.
Discover how Verizon is transforming the DoD, strategically and securely.
1 Defense Advanced Research Projects Agency, "The Evolution of ARPANET," page 2.
The author of this content is a paid contributor for Verizon.