The Defense Department is going all in on artificial intelligence, putting money into research, implementing it on multiple levels and setting up a centralized center to speed up its adoption, however a new report finds that the way the Pentagon is going about its work on AI maybe hurting its efforts.
A study from the RAND Corporation, and mandated by Congress, finds DoD’s posture in AI is significantly challenged across all dimensions.
Insight by Ciena: In this exclusive executive briefing, experts will discuss the wide-area broadband about to go out of this world.
In response to the report, DoD did not pushback on the findings. A statement from DoD Chief Information Officer Dana Deasy from earlier this week says DoD will embrace the results of the study.
“We welcome RAND’s independent assessment of DoD’s posture for artificial intelligence,” Deasy said. “The RAND report provides a thorough and thoughtful critique of the state of AI within DoD. It highlights opportunities and challenges while offering a number of helpful recommendations to accelerate fielding AI-enabled capabilities at scale. We will take this assessment along with other recent reports by the National Security Commission on AI and the Defense Innovation Board to forge a way forward for the multi-generational task of accelerating the fielding of AI-enabled capabilities throughout the U.S. military.”
DoD created an AI strategy in 2018 and set up its Joint Artificial Intelligence Center (JAIC) as a focal point for the department’s AI efforts.
The authors of the study found issues with those crucial entities.
“One of the main findings is the visibility, the authorities and the long-term resource allocation given to the JAIC do not seem to align with what was stated in the AI strategy as this significant mission of the JAIC,” Danielle Tarraf, senior information scientist at RAND, told Federal News Network. “That raises questions about whether the JAIC can really do what it’s been tasked to do.”
The study found that DoD does not have proper metrics to measure the success of the JAIC either.
That issue permeates throughout the military. The military services created their own AI hubs to complement JAIC.
“As with DoD’s AI strategy, the service annexes generally lack baselines and metrics to meaningfully assess progress,” the authors said in the report. “Moreover, although the services have creates centralized AI organizations, the roles, mandates, and authorities of these organizations within the services remain unclear.”
It’s not just the organization and strategy that’s an issue, however. RAND found that the data and talent needed to implement AI and the advancement of AI technology itself are a hindrance to its adoption.
DoD and the military services are making significant pushes to label data, make it usable and make it accessible, but there’s still a long road ahead.
“In order really achieve success in AI you need to lineup not only structure, mission, resources and technology, but talent and data are also equally critical,” Tarraf said.
“DoD faces multiple challenges in data, including the lack of data,” the authors wrote. “When data do exist, impediments to their use include lack of traceability, understandability, access, and interoperability of data collected by different systems.”
DoD is also dealing with a talent issue in technological fields. There aren’t enough AI experts to go around and DoD can’t pay them as well as the private sector.
Finally, the authors note that the current state of AI is nowhere close to ensuring the performance and safety of AI applications where important systems are concerned.
Want to stay up to date with the latest federal news and information from all your devices? Download the revamped Federal News Network app
That may be an issue, considering JAIC wants to use AI in combat missions this coming year.
To remedy the situation, RAND offered 11 recommendations. Those include rebalancing authorities to give JAIC the power it needs, and developing a five-year roadmap that outlines metrics and baselines for assessment.
RAND also recommended frequent portfolio reviews, and workshops to look at investments and the state of AI.