How to Use Artificial Intelligence in Diplomacy

Foreign affairs agencies should take a proactive approach to using the new technology to revolutionize the conduct of international relations.

Alexander Hunt, public affairs officer at the U.S. Embassy in Guinea, visited Raby et les Enfants, an association working to encourage children to take charge of their well-being, in July. Hunt's section has been using AI tools in public diplomacy work since late 2022. Photo by Mobe Studio via Instagram.

Secretary of State Antony Blinken called last week for “fostering an environment” globally where innovation in artificial intelligence (AI) can flourish. “We can deepen our cooperation across governments to… incentivize more affirmative uses for artificial intelligence,” Blinken said in a speech in New York. “As we work to manage risks, we can and we must also maximize the use of AI for the greater good.”

The State Department has relied on AI in the Foreign Service selection process since 2015 to score candidates’ essays and assess their educational and professional qualifications — the “most significant change” to the process since 1930, according to the department. Last year, computer software took over the critical function of deciding whether a candidate advances in the process from a written test, which served as a “pass/fail gateway” for decades, department officials said. Applicants still have to take the test, but their scores are now combined with scores the A.I. gives their applications, based on their qualifications and six essays.

At the same time, however, the department prohibits candidates from using AI in writing their essays, warning that it “will discontinue any individual’s candidacy if found to have violated this prohibition on use of AI tools in the application process.” Earlier this year, the agency even forbade its employees to utilize AI when creating narratives in their annual performance evaluations and in assessing the work of subordinates. It said in a cable to staff that it was “reviewing the security and privacy implications of the use of generative AI.” A State Department official said this week that the department’s policy is “not to use” AI for processing what is known as “sensitive but unclassified information,” which covers employee evaluation and promotion records.

Although it’s prudent to “manage risks,” as Blinken put it, AI can revolutionize the practice of diplomacy for the better — from enhancing communication and analysis to providing innovative solutions to complex challenges to leveling staff capabilities. AI can even improve the department’s evaluation and promotion process, which employees have criticized as antiquated and inequitable for decades. Foreign Service promotions often hinge on the writing skills of supervisors, rather than on employees’ actual accomplishments or performance impact. A boss who is a bad writer can easily cost a subordinate a promotion. Integrating AI into the process could level the playing field. The ban should not last long, especially considering the difficulty of policing AI use among tech-savvy staff and the growing pressures on employees due to increased workloads and staffing shortages worldwide.

As a former Foreign Service officer for three decades and ex-CEO of a technology company specializing in AI and machine learning, I encourage the State Department and other foreign affairs agencies to embrace a broader perspective. They should harness and create AI tools across workloads, including for communication, public diplomacy, writing and analysis, which would improve performance, enhance capacity and make a more significant impact. While the current configuration of these tools may not always be suitable for classified work, it is plausible that an internal or controlled version could be developed in the near future, similar to the availability of spell check and text-editing across classified and unclassified systems. Wherever appropriate, AI tools could initiate drafts at the unclassified level, which can later be transferred to classified systems.

U.S. diplomats are already benefiting from AI at embassies overseas. The public diplomacy section in Guinea, for example, started using ChatGPT, a model developed by the U.S. company OpenAI, in late 2022 to draft daily media summaries for the ambassador and embassy staff. The new technology allows the post’s media specialists to produce a draft “in a matter of minutes,” Alexander Hunt, the section chief, told the U.S. Advisory Commission on Public Diplomacy in June. At embassies not using AI, it can take press specialists, who are usually host-country natives, hours to scour the local media, put together comprehensive summaries and translate them into English.

“We quickly realized that this was a really powerful tool, and the entire section could benefit from it,” Hunt said. “We rolled it out across the section in February, and it has since helped us with everything from drafting speeches and press guidance to crafting project proposals and social media posts. More recently, we’ve actually started exploring other AI tools that can help with image, video and audio — things like helping us with graphics and video clips, clean up audio, edit photos and the like.”

Get the Diplomatic Diary

Hunt said that his staff is trained to use any AI-generated content only as “a first draft, and they have to verify all of the details that ChatGPT puts out.” Hunt’s team has also trained ChatGPT in the format of the embassy’s media summaries. “The better the input, the better the output. We’ve trained our team to be very specific in the prompts that we use. We try to always include context about how the product will be used, who’s the audience, what’s the outcome, length, format, style, tone, etc.,” Hunt said. “We’ve found that if we can give it a much more specific prompt — for example, ‘Write a speech for the U.S. ambassador to Guinea for a diplomatic reception. The audience is mostly women. It should be five minutes long. Include references to prominent female historical figures in Guinea and a quote from a prominent female American figure.”

The technology also builds institutional memory over time and can help newly arrived diplomats with a brief of noteworthy developments in the host-country, so they can better understand the local context, Hunt noted. “We think of ChatGPT as producing the middle 60 percent of the work,” he added. “As a team, we dedicate our time and human brain power to three things: The first is to clearly articulate our needs at the outset and draft a prompt that’s going to give us the best output. The second is then to check for accuracy, style, content, bias. Finally, we add any observations or analysis that we’ve learned from our engagement with contacts in the field that can help color the final product.” AI is “an amplifier of humans,” Hunt pointed out, “not a replacement.”

What other elements of diplomatic work could be amplified by AI? Today’s technologies could assist diplomats with real-time translation of open-source material across languages, so they can understand sentiment in foreign societies faster and better, which would make the impact of programs far greater. This is critical, because a core objective of diplomacy is to understand foreign environments, and the best understanding comes from listening to people in their own language and their own voice. AI systems are already remarkably good at quickly aggregating millions of data points across platforms and languages, and our foreign affairs agencies should do more to leverage this capability. Predictive analytics is another AI tool that can benefit diplomatic work. For example, it can help identify potential sources of conflict and instability before problems escalate. By analyzing historical data, social-media sentiment and other relevant factors, AI can provide early warnings, allowing diplomats to take proactive measures to prevent and resolve conflicts more effectively.

Last month, the Harvard Business School published the results of a study conducted in partnership with the Boston Consulting Group, which offered “field experimental evidence of the effects of AI on knowledge worker productivity and quality.” About 7 percent of the individual contributor-level consultants at the company, or 758 of them, participated in the experiment. “For each one of a set of 18 realistic consulting tasks within the frontier of AI capabilities,” the paper said, consultants using AI completed 12 percent more tasks on average, and did so 25 percent faster, producing 40 percent higher quality, compared to a group of consultants not using AI. The technology helped participants with below-average performance improve by 43 percent, and those above average by 17 percent.

The study made me think about those U.S. diplomats who don’t get promoted because of the low (or lower compared to others) skill level of a supervisor writing a narrative about the officers’ performance. The analytical and writing skills and work content of the study’s tested consultants is fairly similar to those needed to produce an objective and thorough evaluation in the Foreign Service. There may be hurdles to overcome, such as privacy concerns about employee files in an AI system controlled by a private company, but solutions are certainly possible to find. The State Department and other foreign affairs agencies should take note and choose a proactive approach, rather than fall behind the private sector.

Diplomats like Hunt are ahead of the curve, but only institutional change will make a meaningful difference. AI technology “is as consequential for society writ large as the web browser or the smartphone or social media,” Hunt said at the June meeting. “It’s advancing extremely quickly, with major announcements from tech companies almost every day. AI is here to stay, whether we like it or not.”

Applied correctly and used competently, AI tools can power up diplomacy across the board, including the employee evaluation and promotion process.