State dives deep into data

Ryan Dukeman, an analytics branch chief in the State Department, said there is "tremendous potential" for AI tools at the agency.

Ryan Dukeman, an analytics branch chief in the State Department, said there is "tremendous potential" for AI tools at the agency. Courtesy: Ryan Dukeman

The State Department is increasingly capitalizing on advances in data analytics to inform diplomacy and funding efforts.

Information overload is an occupational hazard at the State Department, where analysts and diplomats must make sense — and execute good foreign policy — from a firehose of data pouring in from around the world. With a focus on emerging technology tools and stronger data analysis capabilities, the department is working to make all that information more accessible and useful.

The agency is applying its data operations across multiple tech and policy efforts, ranging from developing better interoperability throughout the enterprise and gauging cybersecurity risk to determining how to allocate resources for diplomacy initiatives. 

“When you think about foreign policy analysis as it is traditionally done, so much of it is about synthesizing large amounts of internal and external textual information and reporting to try and derive new insights for our leadership. We really see tremendous potential for some of the new AI tools to accelerate that,” said Ryan Dukeman, an analytics branch chief in the State Department and Federal 100 award recipient for his work incorporating statistical and machine-learning techniques to track Chinese foreign policy activities worldwide. 

Those efforts included his work in the State Department’s Office of China Coordination, informally known as China House, where Dukeman and his team used analytics to determine how to best apply congressionally mandated foreign assistance funding for diplomatic efforts “to have maximum strategic and policy impact,” he said.

Analytics also show potential for systemizing the mass of data flowing through the State Department and making it more accessible.

“We do a ton of textual analytics and natural language processing to try and make better systematic and archival use of the information we’ve got so that it can be more structured for analysis, more useful to decision makers,” Dukeman said. “[We’re] really trying to sift through the enormous haystack of information.”

For example, the State Department generates thousands of cables every month, he said, and capturing the relevant textual data could help better inform foreign policy analysis.   

Prioritizing data 

Recognizing the importance of the data doesn’t mean much without the tools to capitalize on it. Secretary of State Antony Blinken has made data analytics a department priority, starting with the enterprise data plan’s focus on acquiring new analytics tools and piloting emerging technologies like artificial intelligence and machine learning. 

Dukeman said Blinken’s backing has been critical to the work of the State Department’s Center for Analytics to establish data capabilities in the department’s bureaus. In addition, CFA’s collaboration with the agency's chief information officer has been essential to identifying the technology needed to help inform foreign policy. 

“Secretary Blinken is a huge supporter of data,” he said. “We have kind of an innovative model: our tech vertical within CFA is dual-hatted with the CIO shop in the department. So, we are really synced at the hip about onboarding the infrastructure for advanced analytics.” 

That infrastructure includes cloud capabilities to help securely manage datasets, full stack data generation, access control, secure governance, large-scale analytics and business intelligence to manage the center’s work on things like text analytics, Dukeman said. 

As for using AI/ML to assist in those efforts, the State Department had been monitoring large language models before ChatGPT was mainstreamed, but its interest in adopting and scaling internal AI capabilities has increased.

In April, the department issued its internal policy on AI governance, including the maintenance of its inventory of AI use cases and the roles and responsibilities of officials overseeing AI-enabled uses and other policies. An enterprise-wide AI strategy is in the works as well. 

Meanwhile, the department’s current AI inventory lists an expanding range of pilots that are conducted across the enterprise, ranging from using natural language processes to capture earmarks in appropriations bills to detecting deepfake content.

Hiring the unicorns

Alongside the technology focus is a dedicated strategy to recruit and reskill the State Department workforce to capitalize on data analytics efforts.

“We’re really trying to think about this not just as a tech problem, but as a human capital opportunity,” said Dukeman. “I am really excited [about] … some of these strategic workforce initiatives. When we think about making the State Department a data-informed culture of data-driven decision-making … that’s where the rubber really meets the road.” 

One of the challenges often heard in federal agency circles is how to gain workforce buy-in for new technology applications and build an understanding of their impact on operations. In short, no matter the new tool, it won’t work if employees don’t embrace it. 

“We always think about that — that systems are slower to adapt than people, and there really is an appetite from our leadership at the senior-most level and throughout the workforce, who are the ones who are most experiencing the pain points of legacy IT or non-interoperable systems,” Dukeman said. 

He added that the Center for Analytics is designed to help the State Department’s bureaus develop the data-based work they are pursuing for their own objectives.

“The best advertising is happy customers because they help get new customers and grow the opportunity to bring analytics and data management to bear on other parts of the foreign policy mission that may not have known about our work or not had an appetite for it just yet,” he said. 

Courses on subjects like data literacy and data science are structured for different levels of maturity and focus areas. Data literacy for executives, for example, aims to help department leaders learn how to trust data appropriately, instill a data-informed culture, and address other management challenges. 

Meanwhile, the State Department will continue to recruit data scientists, aided by the Biden administration’s Subject Matter Expert Qualification Assessments (SME-QA). These allow human resources and subject matter experts to prequalify applicants for data science capabilities. The department also will leverage U.S. Digital Corps fellows, bringing in data scientists at the GS-9 level with the hope of retaining them as full employees. 

Dukeman said he’s seeing more and more young professionals pursuing international relations and data science together — a trend the State Department can capitalize on.

“We think of them as unicorns,” he said. “They can understand the foreign policy mission, understand the data science and have the technical chops and have the people skills necessary to do this work in government in a very collaborative way with our partners. And I’m seeing the production of unicorns like that in the wider world really increase, which is [a] joy as a hiring manager.”