What’s DoD Doing to Fill AI Talent Gap?

military dod army AI innovation cyber command

When the Department of Defense considers its sweeping plans for artificial intelligence, the biggest challenges might not be with the technology itself. It just might be with the boots on the ground, even if the ground is inside a data center.

The DoD’s top data officer recently reiterated what CIO Dana Deasy and other officials have also said: that the arc of what the agency can achieve with AI rests on the two primary pillars of good data and AI talent.

Data, which is something the DoD has plenty of, is one part of the challenge. It’s obvious in AI circles that the technology is only as good as the data it’s working with. The data must be comprehensive and structured in a way that algorithms can work with, which seems possible. “The more data you have to train your algorithms, the more accurate the algorithms are and the faster you get your results,” DoD Chief Data Officer Michael Conlin said at the ACT-IAC Artificial Intelligence and Intelligent Automation Forum in Washington, D.C. “So data is everything.”

Talent could be the bigger problem since, as with other in-demand technology areas such as cybersecurity, skilled employees are scarce across the board in industry and academia as well as in government. Data may be everything, but so is manpower. “It’s all about talent,” Deasy told the House Armed Services Committee’s Emerging Threats and Capabilities Subcommittee last month.

In attracting AI talent, DoD faces the usual competitive disadvantages, starting with industry’s ability to pay more, but it also has to overcome something of an image problem with regard to its use of AI and other technologies. In June, Google bailed out of the DoD’s Project Maven, which is trying to use AI to automate the analysis of millions of hours of drone-shot videos, after Google employees protested the company’s participation in a project that would result, eventually, in drone strikes.

One way DoD has courted industry’s involvement in AI development is by inviting companies–from established defense contractors to start-ups–into its development projects. One of the main reasons for creating the Joint Artificial Intelligence Center (JAIC), designed to coordinate work on DoD’s 600-plus AI-related projects, was to promote collaboration with industry. The Defense Innovation Unit (DIU) was launched in 2015 as a way to establish partnerships with industry. The military services also have established their own collaboration projects, such as the Army’s Expeditionary Technology Search.

The department also could quell some of the fears about AI and its potential to act autonomously by focusing on ethics, which has become an emerging topic in the field. The Defense Innovation Board is conducting a review of AI ethics, collecting input on the ethical use of AI. “As we move out,” Brendan McCord, machine learning chief at DIU, said in July, “our focus will include ethics, humanitarian considerations, and long-term and short-term AI safety.” DoD has said its ethical concerns will extend beyond just fears of killer robots to any area where a machine’s decision could affect people, including cybersecurity, civil rights, legal matters, and medicine.

For DoD, ethical use of AI could start in the development phase. A Government Accountability Office report on AI earlier this year included a participant’s proposal for a “computational ethics system” that would seek to instill ethical behavior in the programming. The Defense Advanced Research Projects Agency’s Explainable AI program could also further the ethical use of AI, by finding a means to let AI systems explain, in human terms, the complex logical reasoning behind their conclusions.

Bringing skilled AI employees into DoD is a challenge, just as it has been with cybersecurity, but the department might be seeing some results from its recruitment efforts. JAIC officials recently said they were planning to add 75 employees in addition to contractors.

Whether it’s a matter of addressing ethical concerns, partnering with industry, educating a workforce, recruitment and retention efforts, or some combination of all, DoD has its work cut out for it in filling the AI talent gap.

Recent