If your close combat simulator isn't chatty enough, the UK Ministry of Defense (MoD) has got you covered. The Ministry of Defense has awarded a contract to combat training specialist 4GD to add the ChatGPT language model to its SimStriker robotic targets.
With the development of missiles, drones, robots and long-range sensors, it's easy to imagine the modern soldier sitting in front of a screen all day without ever having personal contact with the enemy. In fact, modern warfare brings infantry and special forces into close contact with enemy forces in urban environments.
Such a war requires high-intensity professional training, because soldiers often have to make split-second decisions based on incomplete information in a very unfamiliar and small space. This means not only completing the mission, but also maintaining the safety of the combat team and being able to distinguish the enemy from civilians and hostages, and distinguish real casualties from ambushes.
This type of training relies on practicing in a variety of scenarios of varying fidelity until the response becomes a matter of reflex and muscle memory rather than a conscious decision. This in turn means that practice environments must sometimes be very realistic.
4GD's SimStriker, which has been in development since 2020, is a system that installs sensors and interactive targets in a maze of corridors and rooms. The facility has so far been used by the British Army's 16 Air Assault Group in Colchester, as well as the Air Assault Group, Essex Police and Ministry of Defense Police.
While SimStriker targets look like high-tech tailoring dummies, they are motorized and equipped with hit sensors that record the accuracy and rate of fire of soldiers in combat. Additionally, they can detect movement, light, and sound. They can also respond verbally, sound an alarm, and fire back with non-lethal weapons.
So far, it's all very realistic, but the Department of Defense hopes to make things even more real by adding OpenAI's ChatGPT system through the Defense and Security Accelerator (DASA) to let targets have so-called "synthetic conversations" with soldiers and their equipment, such as through social media feeds. In this way, training sessions can be more varied, dynamic and immersive, providing more customized scenarios for training supervisors.
James Crowley, Director of Business Development at 4GD, said: "We are very excited about this latest stage of development for SimStriker. Winning this contract from DASA demonstrates that, together with our industry partners, 4GD will continue to adapt our solutions to achieve the best training results. 4GD's flagship SmartFacility product is designed to bring realism to military training simulations, and the use of artificial intelligence adds another level of realism to urban warfare scenarios."