‘Digital Coaches’ Can Produce Expert Soldiers
The ability to deploy, fight and win requires a ready force. Improving today’s Army to accomplish these tasks does not end with modernizing platforms. It continues with improving the skills and decision-making abilities of soldiers.
So soldiers can learn their military occupational skills faster, the Army should invest in developing a “cadre” of artificial intelligence training systems. Unlike human cadre, these machine-learning systems could manifest as virtual personal assistants such as Amazon’s Alexa or Apple’s Siri, or they may be integrated into the sensors and computer hardware in new platforms such as fire control systems in smart weapons that account for human error and provide feedback.
Employing such artificial intelligence (AI) cadre to capture performance measures in real time could not only help individual soldiers, but could also enable broader performance lessons to be distributed across the force. The integration of these “digital coaches,” and the enterprise-level lessons learned they can generate, will improve force readiness.
Moreover, the adoption of such AI trainers has the potential to democratize access to premier-level coaching—and thereby increase development of expert soldiers, the true masters within their fields.
Practice, Practice, Practice
The study of expertise has come a long way over the past couple of decades. Numerous research efforts have examined expert performance and analyzed how champion-level skills are acquired. Researchers have realized that astounding heights of performance can be accomplished by nearly anyone, with the right training regimen and right coach.
Researchers including K. Anders Ericsson, Joseph Baker, Damian Farrow and Addie Johnson have written extensively about development of expertise. From chess grand masters to professional cricket players to concert violinists, common practice behaviors distinguish amateurs from professionals.
What this combined research has found is that the most effective way to develop new skills is through deliberate practice. Ericsson espouses a method involving five components:
- Individuals must aspire to push beyond their comfort zones and seek to improve where they most struggle.
- To do this, they must work toward well-defined, specific goals.
- As they practice a set of skills, they must remain focused intently during practice.
- They also must receive regular, high-quality feedback regarding their performance.
- The result of these efforts will be refined mental models for given tasks. These mental representations enable experts to assess their problems more efficiently, guide their attention to where it is needed most, and allow them to respond rapidly and effectively by initiating ingrained actions—i.e., muscle memory.
What distinguishes experts from novices and amateurs is the quality of the mental representations they develop. Practice the wrong way and bad habits form. Ingrain perfect movements and perfect performance will follow.
In the Army, examples of master warriors exist, from expert marksmen to elite operators. These elite masters, however, are in limited supply and high demand. The use of emerging technologies to expand access to expert coaches increases the likelihood the Army can develop experts at a faster rate.
Through application of advanced training technologies, the Army can democratize access to expertise. The rise of the global internet, and the drive of both individual and institutional educators to share learning via this information conduit, has engendered the development of numerous platforms for self-paced skill acquisition. However, what these platforms bring in access to world-class content they lack in individualized attention and tailored performance feedback.
Individuals require feedback from expert coaches to accelerate skill acquisition and performance improvement. Digital coaches can provide this mentorship. Individuals do not gain as much from grading their own tests. True masters can recognize negative performance patterns and bad habits, and suggest methods to correct these deficiencies. Soldiers, developing skills at all levels, would benefit from coaching delivered from genuine masters rather than learning from instructors typically only one lesson ahead on the syllabus.
Digital coaches are capable of monitoring individual performances, comparing performance patterns with expert performances and providing tailored feedback. Moreover, they can maintain a limitless catalogue of lessons and skills. Most digital coaching systems do so to help with basic educational skills such as reading and arithmetic. For example, psychology and computer science professor John R. Anderson at Carnegie Mellon University in Pittsburgh and his colleagues have shown improved student performance in public schools that adopt adaptive educational tutoring software.
Similar to these artificial intelligence tutors, digital coaching programs can and should be developed by Army trainers to coach skills relevant for soldiers. Beginning with the basics—shoot, move, communicate—digital coaches could be fielded in the form of wearable, deployable monitoring systems (think unmanned aerial vehicles and other follow-along sensor platforms), or systems built into next-generation platforms being developed by the U.S. Army Futures Command. These systems could monitor training and real-world actions, after which or during they could offer feedback and evaluate performance. This data could then be harvested and analyzed across the force to provide higher-level analysis of broad trends and lessons learned.
Today, soldiers in Advanced Individual Training receive the minimal amount of training considered critical to meet their basic occupational specialty needs. This limited training provides a foundation upon which trainers within their gaining units can build. But units often find themselves with limited time to retrain and augment the developmental needs of new soldiers.
A Lifelong Journey
To accelerate the pace of skill acquisition and mastery, the Army should consider employing artificial intelligence training systems as part of the Total Soldier Concept. To help visualize what this might look like, consider rifle marksmanship. Using off-the-shelf technology today, AI firing coaches could be designed to monitor firing performance, update lifelong databases of firing patterns, and provide real-time feedback to shooters and pattern analysis for unit trainers. Soldiers, across their careers, could track their personalized trends, and alter practice techniques and training objectives accordingly.
Imagine recruits entering basic training and beginning a lifelong journey with an AI coach. This digital mentor could accompany them from the range to real-world combat scenarios. Sensor technology exists that can monitor firing from the rifle. This information, fed to a digital coaching marksmanship system, through machine-learning protocols, could identify signatures for poor performance, then coach individuals on how to adjust their actions.
It does not take much imagination to consider similar examples—from battle drills and basic soldier tasks, such as medical response procedures, navigation, etc.—to more complicated tasks involving teams working together. Sensors worn by soldiers, integrated into systems capable of assessing real-world performance metrics and providing tailored feedback, could enable new forms of training and the democratization of expertise.
Another benefit of employing digital coaches would come from the collated data collected across the force. Data harvested from individuals and small units could form the basis for higher-order collections of information. Organizational-level analysis can aid in training-design processes to optimize skill acquisition across the Army. Lessons can and should be learned from digital coaches. Through omnipresent connectivity and cloud-based management of these AI tools, lessons could be uploaded and used in near-real time to update and modify deployed digital coaches, globally.
A Difficult Task
Developing a cadre of artificial intelligence trainers will be no easy task. First, experts who can identify and record the best ways of performing skills across the large portfolio of MOSs must be paired with data scientists capable of engineering corresponding machine-learning systems. These systems then must make their way into hardware and user-friendly interfaces that can withstand a range of training environments. This will require matching the right technical expertise with the right level of understanding of military training requirements.
Additionally, provisioning an enterprise DoD information network architecture capable of hosting digital coaches, and providing lifelong access for soldiers to their coaches and training data, will be challenging. Those charged with implementing such solutions are also required to operate and maintain today’s networks. The priorities of these technical leaders are defined by the Army’s vision for communications and programmed information technology modernization initiatives. Therefore, those with the greatest demands to maintain readiness levels would have to make this effort their priority.
Today, as much of DoD considers applications of artificial intelligence to improve lethality, it has primarily focused on the performance of weapons systems and platforms—both critical objectives in this time of accelerating revolutions in military affairs. However, the use of AI to improve individual and team skills can produce similar benefits to lethality, among many other aspects of work performance that benefit from better-trained soldiers.
The Army must continue to seek innovative ways to maintain the skills of soldiers. By developing a cadre of AI trainers, and enabling lifelong learning through the democratization of access to these digital experts, the Army can accelerate skill acquisition and increase the percentage of soldiers achieving true mastery of their craft.
The time has come to not only invest in AI-enabled platforms, but also to find ways to use advancements in artificial intelligence to improve soldiers’ skills.