Like any new technology, the race to deploy the best has been met with a hesitation to vet and assess. Image: Tada Images/Shutterstock.com
By: Nicolas ChaillanFollowing Generative AI guidance his office released in June, Army CIO Leonel Garciga spoke enthusiastically at TechNet Augusta to the future of generative AI (Gen AI) within the Department of Defense (DoD), stating that:
“I’m super excited in this space right now, because really, industry has come to the table in a very good way with lots of lessons learned, and everyone is kind of being pragmatic, like, ‘Here’s where we are in this story; this isn’t a 100% solution; maybe this gives you some lift.’ I think this next stage is how do we make it available more broadly.”
Converting Gen AI excitement into adoption is paramount. Urgency is driven by the fact that while DoD agencies are moving forward at varying speeds with Gen AI, foreign adversaries have their foot planted firmly on the accelerator to gain every possible competitive advantage.
Extending Gen AI across DoD and more broadly, the public sector, requires an approach oriented around understanding real and perceived challenges, and the consequences for mission and country of restraining adoption.
Gen AI must be available at highest security levels
Historically, Generative AI adoption across DoD has trailed demand as leadership slow-walks adoption while waiting for commercial contractors to meet required levels of security compliance (IL5, FedRAMP High, etc.). Agency leader overreach slammed the brakes on Gen AI by banning personnel from using government equipment (devices), resulting in some users turning to personal devices or accounts instead of a pre-screened and approved agency-sponsored Gen AI solutions, ultimately creating cyber risk. The tide is shifting as secure, authorized solutions enter the market and industry leaders strive for balance between innovation and concerns that come along with adoption of this emerging technology.
In September, we saw a Gen AI platform launch on cARMY, the cloud infrastructure of the U.S. Army, making it the first DoD service to embark on an AI transformation of this scale. While the Army is moving aggressively, it is doing so with strong levels of compliance and security by deploying generative AI in an Impact Level 5 (IL5) environment. This unlocks soldier access to Gen AI on a Controlled Unclassified Information (CUI) fabric, securely air-gapped, to accelerate the velocity of its software, cybersecurity and acquisition teams.
Beyond compliance, a security-first approach must be applied through Gen AI evaluation, deployment and management. Potential weaknesses should be acknowledged, and through processes of red-teaming, advanced monitoring and zero trust, security experts are constantly working to improve system integrity and prevent breaches in deployed tools.
DoD must have confidence to buy, not just build
The U.S. Army is establishing a strong precedent that can be followed DoD-wide, leveraging industry advancements to support tangible and secure use cases. This leads to the second part of the challenge – contractor readiness. DoD and any agency whether to build their own generative AI platform will face cost and complexity headwinds. And let’s not forget about the tremendous sustainment cost for both the Generative AI platform and GPU requirements.
Generative AI can seem easy at first glance for teams, but the reality is that security and the insane pace of Generative AI make it a tremendous challenge at scale. There is a struggle to identify and build teams with the required technical skills, while commercial partners can come in with the technology out of the box, the cybersecurity authorizations and the expertise to move quickly.
To deliver Gen AI solutions, contractors themselves will increasingly need to tap the technology to both streamline internal processes and serve customers. Leveraging authorized, government-ready commercial solutions for Generative AI reduces the time and resources traditionally spent by the Defense Industrial Base to navigate complex regulatory processes rather than focusing on accelerating innovation and time to deployment.
Lock in on high-impact use cases, not hype
In recent comments about Generative AI, Garciga laid out a methodical approach built for success: “We got more people at the gates than we can handle right now. We’re just making sure we set a foundation that makes sense and that one, we can help people make sure that they don’t jump in there, run out of money, and not get what they’re looking for. Two, leverage models that aren’t going to be supportive of their requirements, which will be disappointing. Or three, I think the bigger thing is to start leveraging technology that really doesn’t augment the kind of work that they’re doing. So we’re trying to balance all that right now.”
While not always visible to external audiences, a look under the hood of the operations at several government agencies reveals the transformational impact Gen AI is already delivering to support, such as software development, acquisition, cybersecurity and data analysis.
Generative AI is accelerating the performance of teams across the government. Like any new technology, the race to deploy the best has been met with a hesitation to vet and assess. We are now entering a new phase, where a Gen AI platform is authorized and government leadership is focused on the potential for greater mission delivery.
Nicolas Chaillan is founder of Ask Sage and former Chief Software Officer for the U.S. Air Force and U.S. Space Force
Trump to move Space Command headquarters out of Colorado
National Guard Order Calls for Hiring More Fed LEOs, Military Unit Under Interior, DHS, Justice
Audit: Rise in ‘Severe Staffing Shortages’ at VA Medical Facilities
Marines Seek New Boots on the Ground
Can My Military Discharge be Upgraded? Yes
The Rules for Getting Both Military Retired Pay and a Federal Annuity
Credit for Military Service for Federal Retirement Annuities
