U.S. Army Launches Generative AI Platform in Groundbreaking Move
The U.S. Army is making military history, announcing that the branch is becoming the first-ever U.S. Department of Defense (DoD) service to implement a generative artificial intelligence (AI) platform into its operations.
With the help of Ask Sage, a leading generative AI platform that assists in both the private and public sectors, the Army has deployed the platform enterprise-wide on cARMY, the Army’s general-purpose cloud environment. It has already proved to help officials in several areas. Army personnel can now code faster and analyze complex data more efficiently, according to Nic Chaillan, CEO of Ask Sage.
Developers are now able to code 35 times more quickly. Acquisition experts can now complete 50 tasks in the same time it previously took to complete one. And officials can analyze thousands of documents in minutes rather than days, according to Chaillan.
Chaillan calls this innovation and integration crucial to staying ahead of U.S. adversaries and minimizing threats from other countries.
“China has deployed Baidu GPT all the way to their top-secret fabric,” Chaillan said. “They already have more people, and they are pretty nimble, and they have generative AI, so now if you are going to keep up and not get the nation at risk, you need that in the U.S. as well.”
The Army began to experiment with Ask Sage’s generative AI platform about six months ago. Ask Sage said in a press release that it started by opening 400 new user accounts with a goal of growing to an enterprise-wide deployment. Chaillan calls it a success.
“We’re not playing and testing things anymore,” Chaillan said. “We are good to go. This is an enterprise production-ready capability, and so that is pretty exciting, and it is already spreading across the department.”
As for Army personnel, leaders do not expect this improved technology to take any importance away from human beings. Leonel Garciga, the Army chief information officer, said he thinks it will be the opposite.
“On the contrary, I think what we are really finding is that we are enabling more capacity for the workforce we have. I could only do 10 of these before. I can now do 100, but I have had a backlog of 1,000 for the last 10 years. We are finally getting to a point where we can do our mission,” Garciga said.
In fact, Garciga went on to say that humans are better at identifying mistakes than the technology.
“When we do some of these polls, whether it is a hallucination or maybe the base data is not right, I am telling you the attorneys are really good at catching [mistakes],” Garciga said. “The contract officers are really good. They are like, 'Yeah, that is not right.' So, I think that helps us because we refine the models, so they get better.”
And during a discussion with SIGNAL Media, Garciga said Army officials have already seen an example of that type of mistake.
“Some stuff we are doing with PEO [Program Executive Office] Aviation on aviation safety, kind of seeing that the initial tranche of data we had that was from kind of open-source space,” Garciga said. “We had a lot of great data from NASA, which was not very relevant to an Apache. So, we had to really rebuild that model and focus our very specific data around military aviation safety.”
The Army is the first DoD service to integrate generative AI enterprise-wide, but Chaillan said they are working with the Air Force, Space Force and parts of the Navy.
“It is more like different teams using the platform,” Chaillan said. “We have about 20,000 airmen and guardians on the platform, which is pretty awesome. And then we have NAVAIR [Naval Air Systems Command] in the Navy. It is not a Navy-wide capability; it is just NAVAIR. So again, limited scope compared to the Army-wide deployment, but it is the beginning. People are growing rapidly, so we hope that the Army leading the way will lead to others also jumping on the wagon.”
