Japanese Defense Minister Shinjiro Koizumi, left center, visits the office of Palantir Technologies Inc. in Washington, Jan. 16, 2026, in this photo posted on his X account.
TOKYO — Artificial intelligence (AI) has increasingly revealed its threatening side amid a series of cases in which the technology has been misused in cyberattacks on companies. At the same time, its use is expanding in the military sphere. How can its risks be managed while harnessing its advantages? The Mainichi Shimbun examines the current state of AI and national security.
On Jan. 16, at the Washington office of U.S. data analytics giant Palantir Technologies Inc., Japanese Defense Minister Shinjiro Koizumi could not hide his enthusiasm.
“This is really good,” he said.
Koizumi had just watched a demonstration video of the Maven Smart System (MSS), an AI-powered command-and-control platform that Palantir provides to the U.S. military to support its operations.
Japanese Defense Minister Shinjiro Koizumi, right, visits an office of Palantir Technologies Inc. in the United States, Jan. 16, 2026, in this photo posted on the X account of Mike Gallagher, who heads Palantir’s defense division.
The visit to Palantir was one of the key items on Koizumi’s itinerary during his trip to the United States for talks with U.S. Defense Secretary Pete Hegseth, reflecting Koizumi’s desire to “learn about new ways of warfare.”
Koizumi was welcomed by Mike Gallagher, a former member of the U.S. House of Representatives who heads Palantir’s defense division, and listened to explanations about how AI is being used in both the military and manufacturing.
The previous day, following his meeting with Hegseth, Koizumi had also received a briefing from Pentagon officials on the use of AI. For about an hour, he heard detailed explanations on how the technology could streamline operations, speed up commanders’ decision-making during combat and help their staff draw up operational plans more quickly.
“What is important in ensuring that AI permeates an organization?” a member of the Japanese delegation asked during the briefing.
A U.S. participant cited the importance of leadership. Koizumi later told members of his entourage, “I learned a lot. I’m glad we asked,” according to sources familiar with the exchange.
Japanese Defense Minister Shinjiro Koizumi, second from front right, and U.S. Defense Secretary Pete Hegseth, second from front left, attend Japan-U.S. defense ministerial talks at the U.S. Department of Defense, Jan. 15, 2026, in this photo posted on the Japanese Defense Ministry’s X account.
On Jan. 26, about a week after returning from the United States, Koizumi gathered civilian and uniformed senior officials and other personnel responsible for AI, cybersecurity and related fields in the minister’s office at the Defense Ministry in Tokyo’s Ichigaya district.
The purpose was to launch an “AI adoption promotion team” under his direct supervision. In a post on X, Koizumi stressed that Japan needed to step up its response as AI transforms national security and the nature of warfare.
Koizumi had shown an interest in using AI even before the trip, but one individual close to him recalled, “His visit to the United States was probably what prompted him to get serious about it.”
Palantir’s MSS was reportedly used in a U.S. military raid on Venezuela in January in which President Nicolas Maduro and his wife were captured in just two hours. The system was also reportedly used extensively for target selection when the United States and Israel attacked Iran the following month.
Already used in exercises with US
Japanese Defense Minister Shinjiro Koizumi, left, checks preparations for the Japan-U.S. joint exercise Keen Edge alongside Hiroaki Uchikura, chief of staff of the Joint Staff, the top uniformed officer of the Japanese Self-Defense Forces, at the Defense Ministry in Tokyo on Jan. 26, 2026, in this photo posted on the X account of the Joint Staff Office.
MSS was also used in a joint command-post exercise involving Japan, the United States and Australia that was held between August and September.
The Japanese Defense Ministry’s budget request for the next fiscal year, announced in August, included a project to establish a decision-support system for the Japan Joint Operations Command by introducing AI into command-and-control systems for Self-Defense Forces (SDF) units.
The ministry plans to accelerate the system’s development using existing foreign technologies, including those from the United States, while strengthening homegrown AI technology so that foreign and domestic systems can be used together. MSS is seen as one of the leading candidates for adoption.
Japanese Defense Minister Shinjiro Koizumi, left, and U.S. Defense Secretary Pete Hegseth attend Japan-U.S. defense ministerial talks at the Pentagon, Jan. 15, 2026, in this photo posted on the Japanese Defense Ministry’s X account.
Within the Defense Ministry, however, concerns have been raised about how errors caused by AI should be handled and who should be held responsible. There are also reservations about entrusting foreign-made AI with sensitive information related to the command of SDF units.
(Japanese original by Nozomu Takeuchi, Political News Department; and Yoshitake Matsuura, Editorial Division)