An artificially intelligent tutoring system that semi-automates team training methods has been successfully demonstrated in a Navy laboratory and will be studied further at the AEGIS Training and Readiness Center in Dahlgren, Va. The center teaches individuals to work together as an integrated, high-functioning team. Although the center's methods are highly effective, the exercises require a one-to-one ratio of instructors to trainees. The artificially intelligent tutor developed by ONR-funded researchers semi-automates the training methods to reduce the number of instructors needed. The system encodes and interprets keystroke and trackball actions, eliminating the need for a human observer to do this. Its also tracks eye movement to determine whether the trainee is looking at the right information at the right time -- something a human observer located behind the trainee has difficulty judging. Automatic speech recognition determines whether the trainee is communicating the right information at the right time. The system offers prompts and immediate feedback on performance throughout the exercise. A human instructor gives a pre-briefing, makes judgements of high-level team behavior, and conducts debriefing discussions. Instructors use handheld computers during training to access the system. Critical incidents in team behavior can be captured and replayed during debriefing or any other time.