AI commission tells DOD to get AI infrastructure ready by 2025, or else

In a draft of its final report to Congress, the National Security Commission on AI continues to hammer home the importance of laying a foundation of AI development in DOD.
Eric Schmidt, Defense Innovation Board (DIB)
Chairman Eric Schmidt speaks May 3, 2019, at a Defense Innovation Board event at the Pentagon. (DOD / Army Sgt. Amber I. Smith)

The Department of Defense needs to widely implement an infrastructure to support artificial intelligence by 2025, an independent advisory committee recommended in a draft of its final report to Congress.

The National Security Commission on AI, chaired by former DOD Deputy Secretary Bob Work and Google CEO Eric Schmidt, says that 2025 is the year that basic infrastructure needs to be in place for the military to widely adopt the family of emerging technologies. Accomplishing that will depend heavily on IT modernization and better harnessing the department’s data, according to the report.

Once that AI infrastructure is in place, the Pentagon will have a chance to work toward some of its greater objectives, like building AI-enabled, sensor-based command and control networks.

But if the department doesn’t get to such a place, there is little to no chance it will achieve the ambitious AI-based goals it has for things like Joint All-Domain Command and Control and other programs that will help it keep pace with — or prevent falling further behind — in the AI race with modern adversaries.


“The Department of Defense (DoD) must set an ambitious goal. By 2025, the foundations for widespread integration of AI across DoD must be in place,” the draft report states. “Those foundations include a common digital infrastructure that is accessible to internal AI development teams and critical industry partners, a technically literate workforce, and modern AI-enabled business practices that improve efficiency.”

The commission has been a powerful voice in DOD’s AI work so far. Many of the commissioners are former DOD officials, like Work and Katherina McFarland, a former assistant secretary for acquisition. While none of the advice it gives is binding, the commission has had many of its recommendations end up in law.

The report details other digital ecosystem initiatives — such as acquisition reform, pooling AI resources and increased bandwidth to share data — that also must take place for the DOD to reach its goal of infusing AI across the department.

The commission has transmitted several previous reports to Congress with policy recommendations. In those, it has urged Congress to create a digital corps akin to the National Guard and a military cyber academy, and increase the federal research and development budget several times over.

This final report underscores many of those recommendations and adds new ones. Other suggestions include: rejecting bans to putting AI in lethal weapons systems, establishing software teams in each combatant command and making an academic center of excellence to concentrate university work on defense AI. Many of the other recommendations focus on improving government tech talent, both with new recruiting practices and retraining current employees to understand AI.


If government agencies do not have enough of the right talent, every AI initiative will struggle and most will fail,” Jose-Marie Griffiths, a commissioner who leads efforts on workforce reform, said during a public meeting on the report. 

Latest Podcasts