The Foreign Service Journal, October 2022

THE FOREIGN SERVICE JOURNAL | OCTOBER 2022 13 favor and advance one another’s careers through the assignments process and while sitting on promotion panels. Bruce K. Byers FSO, retired Reston, Virginia AI and the FS Selection Process I write this letter to the editor of the FSJ after reading an article by Charlie Keohan and Nicholas Kralev on the Diplomatic Diary blog of the Washing- ton International Diplomatic Academy, headed by Kralev. They write that in April the State Department announced what it described as “the most significant change in Foreign Service entrance procedures since 1930.” State made an artificial intelligence (AI) algorithm test, including textual analysis of six essays by an applicant, the co-dominant if not dominant factor along with the traditional Foreign Service Exam (FSOT) in determining who would move on to an oral exam. Previously, the FSOT had been the determining factor in screening at the first stage. This change was announced without prior consultation with AFSA, which appears to have been a deliberate eva- sion since the importance of the change (and thus the consultation requirement) could not have been unknown to those who implemented it. Beyond that exercise in bad faith, the new procedure raises several troubling questions for the future of recruitment and thus of the Foreign Service itself. This is particularly the case because there has been no transparency by the department regarding how those algo- rithms work and, above all, what factors are favored and how they are rated. 1. The test score will be based, in part, on the applicant’s “personal qualifica- tions.” What does that mean? Could this be a device for introducing racial or ethnic quotas or biases into recruitment? Note that Deputy Secretary for Management and Resources Brian McKeon in May testi- fied to the Senate Foreign Relations Com- mittee that the changes were aimed at enhancing “diversity” in the Service. That code word suggests possible racialism in recruitment abetted by a supposedly neutral algorithm. 2. The algorithm will also analyze six essays written by appli- cants to evaluate their qualities. Accurately evaluating a human’s qualities by a computer’s statistical algorithm of their writings strikes me as fanciful. And all the more so because those six essays will have been written by applicants with ample time to try to “write to the test.” In the “old days,” the applicant was given a choice of three topics and chose one on the spot on which to write an essay within 30 minutes. The panel then evaluated it. I acknowledge that this may not have revealed the inner personality or many qualities of the applicant, but at least it honestly tested their mental quickness on their feet and ability to write, which are essential qualities for the Foreign Service. So why do we now need six essays instead of one from an individual? That number six is apparently an excuse for use of AI computers instead of analysis by human panels who, it is said, would be inundated by essays. But if a person cannot express and reveal themselves adequately in one essay, why six? Foreign Service officers have for decades lamented the lack of transpar- ency in promotions and assignments. These latest changes add to that a lack of transparency in recruitment, on which the future of the Service depends. It strikes me as either the fruit of techno- cratic foolishness or an attempt at racial and ethnic bias in recruitment. Marc E. Nicholson FSO, retired Washington, D.C. n Share your thoughts about this month’s issue. Submit letters to the editor: journal@afsa.org CORRECTIONS The September FSJ letter from the editor incorrectly pointed to Secretary Blinken’s August visit to Africa as his first visit there as Secretary. He has in fact visited Africa more than once, his first time as Secretary in November 2021. In Gregory Garland’s September FSJ article, on page 47, the last ref- erence to Sékou Touré should read Guinea’s leader, not Ghana’s. We regret the errors.