Artificial intelligence (AI) is infiltrating nearly every industry and continues to be a hot topic in the training industry. Could AI for military training improve training effectiveness? Earlier on Modern Military Training, we explored the possibilities of AI for training as well as the challenges it can present when organizations push forward without understanding requirements and desired outcomes. Yet, there are many industries that have put in place several best practices from which the military training field can learn.

Take the healthcare industry, for example. Healthcare professionals and data researchers are using data analytic capabilities for genome research, to deliver more accurate, faster results, to enable the full potential of telehealth. Through many of their efforts, the healthcare industry has provided a model that could be transferable to the military training industry.

According to the American Health Information Management Association, the Five Rights Model for Clinical Decision Support (CDS) requires that following: the right information, the right person, the right CDS intervention format, through the right channel and at the right time in workflow. When designing an adaptive learning system, this same model can be applied.

Let’s look at it more closely to see how those same five rights can be applied to the design for an intelligent adaptive learning system:

1. The right information. Determining the right information is a complex process. It begins with defining the problem, then choosing the right machine learning technique to solve the problem (i.e., classification, estimation, identification, pattern recognition or exploration). And finally ensuring the right data and the amount of data has been collected to support the process.
2. The right person. One of the advantages of adaptive learning is to personalize the training to meet the needs of each individual student. To ensure training effectiveness, measuring performance tasks alone is not sufficient. The system must have the capability to assess each student’s task performance and cognitive state to adapt the learning to maintain an optimal level of engagement during the training exercise. The system must also be able to aggregate performance assessment to evaluate teams.
3. The right format. The adaptive learning design must consider how feedback is given (i.e., human instructor, intelligent tutor, or a combination of both). Feedback is a critical success factor in the learning process for knowledge and skill retention. There are many advantages to utilizing a human instructor to provide feedback to the student. For example, they can read and interpret cues in human behavior that are linked to affective states associated with cognitive performance. There is an ongoing effort to decrease significant costs due to time demands on the human instructors monitoring performance and providing individualized feedback in the training industry. Intelligent tutoring systems are being developed to enable autonomous learning environments.

There are several methods that could be employed to adapt training, such as displaying a performance grade or data visualizations on an instructor dashboard for after action reviews, providing an alert to the student on a Heads-Up Display (HUD), or through haptic feedback of the training device. Some may argue that a blended, human-machine approach is the best model.

4. The right channel. Another adaptive learning design consideration is the learning progression. When the system makes a recommendation for further instruction, a learning proficiency model should be referenced, outlining a learning continuum or roadmap of where students are trying to go. These models are typically application specific.
5. The right time. With the emergence of real-time data collection and analysis, determining when the appropriate time is to deliver feedback can impact the transfer of requisite skills. Current models rely heavily on summative assessments, evaluating student performance through a grade or score at the end of a course. This is a one-size-fits-all approach. Formative assessment methods utilize new technologies with real-time analysis capabilities, providing an integrative evaluation approach for students to progress at their own pace. This allows more proficient students to progress quickly through the material and provides remedial training to those who need additional assistance.

While there is still significant research and development to be done before AI in military training is sustainable and effective, learning from others in the industry allows us to move faster and incorporate proven best practices gets us there faster.