The Pentagon\u2019s Chief Digital and Artificial Intelligence Office has been in a period of transition since the department folded it into the Office of the Under Secretary for Research and Engineering last year. Previously, CDAO reported directly to the deputy secretary of Defense, and some analysts warned the move could sideline or bury AI efforts within the organization.<\/p>\n
But Andrew Mapes, acting principal deputy chief digital and AI officer, said the move is doing the opposite: It is positioning CDAO to move quickly to deliver AI capabilities at scale across the department.<\/p>\n
\u201cWith the continued unification of the innovation ecosystem under the chief technology officer, I think there\u2019s been a great opportunity for us to just continue to move much more quickly. The realignment within the Chief Technology Office has enabled us to be much more focused on execution of capability and really working with stakeholders to identify what those enterprise needs are for the department,\u201d Mapes said during Federal News Network\u2019s AI & Data Exchange 2026<\/a>.<\/p>\n \u201cSmall things like not having to maintain an enterprise business office just for our own organization and really getting us laser-focused on delivering capability to the warfighter \u2014 that\u2019s, I think, for me, one of the biggest positive changes that we\u2019ve had since we\u2019ve realigned within the CTO. I know there was a concern about moving AI down. But if anything, I think it\u2019s made it more prominent, and it\u2019s freed us up to move much more quickly to be able to execute the AI strategy.\u201d<\/p>\n Looking to reduce duplicative tech efforts<\/p>\n Last year, Under Secretary for Research and Engineering Emil Michael consolidated several innovation organizations \u2014 including the Defense Innovation Unit \u2014 arguing the Pentagon\u2019s innovation ecosystem had become fragmented, with overlapping responsibilities and unclear authority. Michael also narrowed the Pentagon\u2019s list of \u201ccritical technology areas\u201d from 14 to six, including AI, contested logistics, quantum, hypersonics, biomanufacturing and directed energy.<\/p>\n Mapes said elevating applied AI to the forefront \u2014 along with CDAO\u2019s realignment under Michael\u2019s office \u2014 sends \u201cstrong signals to the department\u201d that the technology is now a clear priority in ways \u201cwe may not have had in our prior alignment.\u201d<\/p>\n The pivot has refocused the CDAO team on executing Defense Secretary Pete Hegseth\u2019s AI strategy, as well as the national AI action plan, Mapes said.<\/p>\n \u00a0\u2018Pace-setting projects\u2019<\/p>\n The AI strategy<\/a>, released in January, tasked CDAO with spearheading seven \u201cpace-setting projects.\u201d The office has already rolled out one of those efforts, GenAI.mil, a platform designed to bring generative AI tools to the department\u2019s personnel.<\/p>\n While the initial rollout featured Google\u2019s Gemini model, which can be used for sensitive but unclassified data, the department plans to introduce additional models that will be available at all classification levels. Mapes said additional models will come online within the next few months.<\/p>\n Other projects include efforts such as Swarm Forge, which focuses on human-machine teaming; Agent Network, aimed at embedding AI agents into command-and-control systems like the Maven Smart System; and Ender\u2019s Foundry, which is designed to accelerate AI-enabled simulation and testing.<\/p>\n CDAO recently issued a solicitation<\/a> inviting technology companies to submit proposals for the Swarm Forge project.<\/p>\n \u201cWe\u2019re going to continue to have opportunities, whether it\u2019s autonomy, whether it\u2019s AI capabilities, multiple different opportunities that are going to be well-suited for performers that may not traditionally have played [together] within the Pentagon. I think it\u2019s a really exciting time to be working with not just CDAO, but across the department. And we\u2019re going to need multiple different players to help us work on some novel and sophisticated issues,\u201d Mapes said.<\/p>\n On the intelligence side, initiatives like Open Arsenal will speed up how intelligence is turned into operational capabilities, and Project Grant will \u201cenable transformation of deterrence from static postures and speculation to dynamic pressure with interpretable results,\u201d he said.<\/p>\n On the business side, the Enterprise Agents project will focus on AI agent development and deployment for enterprise workflows.<\/p>\n Maintaining quality datasets<\/p>\n \u201cNone of the AI capabilities that we\u2019re deploying today are going to be successful if we don\u2019t have our own data house in order. That\u2019s going to be something that we\u2019re continuing to push as far as just the basic fundamentals of data management and ensuring that the war data platform is able to enable, or is able to support, the AI capabilities that the department is continuing to put online, but also more forward looking and partnering with other stakeholders within the department, is the AI infrastructure necessary to support the growth and expansion that is forecast in the future,\u201d Mapes said.<\/p>\n CDAO\u2019s Advana, the department\u2019s biggest data platform for advanced analytics, which was originally developed to support the comptroller\u2019s office, has grown exponentially over the years. It became<\/a> a \u201cvictim of its own success,\u201d prompting the department to restructure<\/a> Advana \u2014 now the War Data Platform \u2014 into three distinct program components.<\/p>\n \u201cAdvana historically has had many purposes and multiple users and a lot of tools and applications that may not have necessarily truly served enterprise purposes. Not to say that they weren\u2019t without value or utility for users but looking to ensure that we are focused and disciplined on ensuring that what we are providing for the department are truly scaled solutions that benefit the entire enterprise,\u201d Mapes said. \u201cFor us, to really achieve the full effects that we feel like CDAO needs to it\u2019s really focusing on ensuring that we\u2019re modernizing the data layer.\u201d<\/p>\n