Government considers humanoid robots for defense, including AI-powered base security and transport

Sankei News reported on September 22, 2026, that the government is considering research into using humanoid robots in the defense sector, including a proposal to equip them with AI and use them for base security and transport. The government says it intends to reduce staffing needs.

If the use of humanoid robots in the defense sector becomes a subject of consideration, a key issue will be how much autonomy and what level of safety the government’s policy of reducing staffing needs in security and transport will allow in practice. The operating conditions within controlled areas such as bases must be made clear; the issue cannot be assessed as an extension of civilian-use robots alone, nor solely on whether AI is installed.

The concept of using humanoid robots in the defense sector is not simply about introducing a new type of machine; it raises the question of whether on-site tasks such as security and transport can be shifted away from human workers. Based on the article, the government is considering the idea at the research stage and has not decided to develop or deploy the robots as equipment. Accordingly, the key points at this stage are less about the robots’ capabilities themselves than about how much autonomy they would have within bases, how AI would be used and how much human oversight would remain. In connection with our previous reporting, humanoid robots have often been discussed in the context of expanding their applications beyond military use to logistics, facility management and factory work. For example, past report 1 and past report 2 showed how robots’ roles are expanding from simple transport to peripheral tasks. This report places the defense sector on that trajectory, but the required standards are higher than for civilian applications because misrecognition or a loss of communications is less likely to be tolerated. See the original source for details. From an industry perspective, these applications involve not only the robot itself but also sensors, recognition software, remote monitoring, communications, power management and maintenance systems. Even based on publicly available information, humanoid robots are integrated systems in ways that go beyond their appearance. If they are to handle both security and transport, requirements are likely to differ by application, including indoor and outdoor mobility, handling steps, nighttime operation and load capacity. In defense applications, the burden of explaining safety is also greater, and the design of demonstrations can readily affect whether the robots are adopted. Points to watch going forward include, first, whether the research involves prototypes or also covers operational testing in the field; second, whether base security or transport is the initial intended application; and third, whether the AI functions are focused on remote assistance or include autonomous decision-making. In all cases, conclusions that go beyond what is stated in the original source should be avoided.

Even if the intersection of defense and robotics expands in Japan, questions will arise over how much of the capabilities needed for actual deployment—including perception, remote monitoring, communications and maintenance—can be handled domestically. Publicly available information indicates that this report points only to a direction of reducing staffing needs, so the specific impact on Japanese component, controls and systems integration companies remains unclear.