As computer gaming technology continues to improve it has come to rival or surpass the simulated imagery, dynamics, and human behavior representation available in current military training simulators. With the goal of applying gaming technologies to training simulations, the Technical Support Working Group (TSWG), through the U.S. Army Research, Development and Engineering Command (RDECOM) Simulation and Training Technology Center (STTC), has sponsored an effort to use a commercial game engine for the simulation of fully automated and adaptive individual adversaries. This paper discusses the use of gaming technology to implement fully automated and adaptive adversarial behaviors. The use of an AI gaming engine allows the adversarial behaviors to adapt by assessing local conditions and dynamically changing tactics, target selection and routing. Learning takes place, and tactics improve, during scenario execution and this learning is retained across scenario runs so that an adversary will improve each time a scenario is run. AI.implant, a commercial artificial intelligence game engine, was interfaced with the behavioral architecture of OneSAF Test Bed (OTBSAF) to provide OTBSAF simulated entities with adaptive adversarial behaviors. Several behaviors were implemented, including a Suicide Bomber and an IED Ambush. A graphical user interface was developed that provides the non-programmer the ability to modify existing behaviors or to create entirely new behaviors. An evaluation was conducted to assess the effectiveness of the AI.implant behavior engine in terms of variability, adaptability and autonomy. Additionally subject matter experts (SMEs) were used to evaluate the ease with which a non-programmer can create or modify adaptive behaviors. The results of these evaluations are provided and discussed. The developed adversarial behavioral system is currently in a usable state, and work to interface the system with existing training systems is discussed.