To prepare for the future, DoD needs to train differently today
Russell Shilling, Ph.D., the senior vice president for government solutions and policy at Riiid Labs, makes the case for why the Defense Department should create a...
Training and education needs to come to the forefront of military priorities, according to the commandant of the Marine Corps, chief of naval operations, and other senior military officials who spoke in December at the Interservice/Industry Training, Simulation and Education Conference, the Department of Defense’s premier event for training technologies. The central message was that training systems need to simulate the way both our allies and our adversaries think to better prepare for future crises.
To accomplish this, the Department of Defense should formally establish education technology certification requirements that apply across the entire department, but also allow the services to have their own service-specific or position-specific supplemental requirements.
Recommendations for technology reform have been building quickly in the DoD. Earlier this year, the National Security Commission on Artificial Intelligence (NSCAI) issued its roadmap for getting the U.S. military AI-ready by 2025. The stakes are high. We are in a hot competition with China for military dominance using advanced AI technologies. In addition to integrating AI systems into our military capabilities, the NSCAI report makes clear that we need to train our government workforce to be AI literate. Only then can we hope for the new technology to permeate our national security culture.
Education and training is, therefore, key. And one of the most effective and efficient ways to educate the workforce in AI is to use AI-based tutors. This approach is key to achieving a holy grail in education: solving the Bloom 2-Sigma problem. In 1984, Benjamin Bloom observed that 1:1 tutoring could raise student scores to the 98th percentile compared to students with regular classroom instruction. The 1:1 adaptive AI tutoring systems at scale may meet or exceed this goal, especially if we incorporate real-time assessment to guide students and tailor the education experience to their specific needs
Past government programs have demonstrated the potential impact of these systems. In 2011 and 2012, the Navy used an AI system developed with funding from the Defense Department’s Defense Advanced Research Projects Agency (DARPA) for training Navy information systems technicians in duty station settings. The DARPA digital tutor trained nearly 1,000 high-school graduates to the level of mid-career technicians in less than half of the time it took with traditional courses. A 2015 study by the University of Michigan and the Institute for Defense Analyses evaluated 50 AI tutoring systems and found that AI tutors generally produced stronger results than other forms of tutoring.
AI capabilities have advanced considerably since then. Deep learning, based on neural networks, has revolutionized the field. We know more about how the brain learns every day, about memory decay and spaced repetition in learning, about the mechanisms that convert short-term memory to long-term memory. Much of that knowledge is finding its way into algorithms.
Today, machine-learning algorithms can track student behavior, trace student knowledge, select the best content for each student to study at any given time and switch gears when the student is bored or frustrated and about to give up.
Combined together in AI tutoring systems, these algorithms can predict with startling accuracy student test scores, a moving number that serves as a sort of carrot: the more students follow the AI tutor’s recommendations, the higher the predicted score will climb.
These systems are domain agnostic and could be used as a standard template for educating the military’s digital workforce. Tech education is not only for officers that develop military doctrine and operating concepts, it should also reach deep into the force structure to operational and tactical levels within the military services, while improving equity throughout
AI and other emerging technologies are changing fast. The state of AI 10 years ago, versus the state of AI today, are very, very different. Hardware has changed, access to data has changed and the algorithms have changed. Certainly, 10 years from now, things will be different still. We need to continue to invest and innovate while developing effective AI-based education tools to prepare both our military and civilians for our future and our national security.
Russell Shilling, Ph.D. is the senior vice president for government solutions and policy at Riiid Labs, a global leader in AI solutions for education.
To prepare for the future, DoD needs to train differently today
Russell Shilling, Ph.D., the senior vice president for government solutions and policy at Riiid Labs, makes the case for why the Defense Department should create a...
Training and education needs to come to the forefront of military priorities, according to the commandant of the Marine Corps, chief of naval operations, and other senior military officials who spoke in December at the Interservice/Industry Training, Simulation and Education Conference, the Department of Defense’s premier event for training technologies. The central message was that training systems need to simulate the way both our allies and our adversaries think to better prepare for future crises.
To accomplish this, the Department of Defense should formally establish education technology certification requirements that apply across the entire department, but also allow the services to have their own service-specific or position-specific supplemental requirements.
Recommendations for technology reform have been building quickly in the DoD. Earlier this year, the National Security Commission on Artificial Intelligence (NSCAI) issued its roadmap for getting the U.S. military AI-ready by 2025. The stakes are high. We are in a hot competition with China for military dominance using advanced AI technologies. In addition to integrating AI systems into our military capabilities, the NSCAI report makes clear that we need to train our government workforce to be AI literate. Only then can we hope for the new technology to permeate our national security culture.
Education and training is, therefore, key. And one of the most effective and efficient ways to educate the workforce in AI is to use AI-based tutors. This approach is key to achieving a holy grail in education: solving the Bloom 2-Sigma problem. In 1984, Benjamin Bloom observed that 1:1 tutoring could raise student scores to the 98th percentile compared to students with regular classroom instruction. The 1:1 adaptive AI tutoring systems at scale may meet or exceed this goal, especially if we incorporate real-time assessment to guide students and tailor the education experience to their specific needs
Get tips and tactics to make informed IT and professional services buys across government in our Small Business Guide.
Past government programs have demonstrated the potential impact of these systems. In 2011 and 2012, the Navy used an AI system developed with funding from the Defense Department’s Defense Advanced Research Projects Agency (DARPA) for training Navy information systems technicians in duty station settings. The DARPA digital tutor trained nearly 1,000 high-school graduates to the level of mid-career technicians in less than half of the time it took with traditional courses. A 2015 study by the University of Michigan and the Institute for Defense Analyses evaluated 50 AI tutoring systems and found that AI tutors generally produced stronger results than other forms of tutoring.
AI capabilities have advanced considerably since then. Deep learning, based on neural networks, has revolutionized the field. We know more about how the brain learns every day, about memory decay and spaced repetition in learning, about the mechanisms that convert short-term memory to long-term memory. Much of that knowledge is finding its way into algorithms.
Today, machine-learning algorithms can track student behavior, trace student knowledge, select the best content for each student to study at any given time and switch gears when the student is bored or frustrated and about to give up.
Combined together in AI tutoring systems, these algorithms can predict with startling accuracy student test scores, a moving number that serves as a sort of carrot: the more students follow the AI tutor’s recommendations, the higher the predicted score will climb.
These systems are domain agnostic and could be used as a standard template for educating the military’s digital workforce. Tech education is not only for officers that develop military doctrine and operating concepts, it should also reach deep into the force structure to operational and tactical levels within the military services, while improving equity throughout
AI and other emerging technologies are changing fast. The state of AI 10 years ago, versus the state of AI today, are very, very different. Hardware has changed, access to data has changed and the algorithms have changed. Certainly, 10 years from now, things will be different still. We need to continue to invest and innovate while developing effective AI-based education tools to prepare both our military and civilians for our future and our national security.
Russell Shilling, Ph.D. is the senior vice president for government solutions and policy at Riiid Labs, a global leader in AI solutions for education.
Read more: Commentary
Copyright © 2024 Federal News Network. All rights reserved. This website is not intended for users located within the European Economic Area.
Related Stories
Navy, Air Force take steps to weed out excessive computer-based training
Air Force meeting recruitment goals, changing mindset around work and training
Army wants to change its cyber training to beef up ranks