Advances in artificial intelligence (AI) and machine learning (ML) are already demonstrating great utility in a variety of domains that point to great opportunities for adoption in a wide variety of military applications. For this reason, data will become the life-blood of many AI and ML enabled technologies. Yet, developing trusted data sets for the purposes of training and testing AI and ML applications will become a central issue to eliciting predictable behaviors of such technology in order to foster trust in autonomous systems. Therefore, the US Department of Defense (DoD) cannot sit idle while AI and ML developments remain forthcoming. Rather, the US military can begin laying the ground work for constructing training data in domain specific ways that will help convergence with AI and ML in the future.

Over the years, Carnegie Mellon University has upgraded its famous snakebot to give it the ability to do things like climb sand dunes and grasp objects. With its latest iteration, you can now add swimming to that list. Work on the Hardened Underwater Modular Robot Snake (HUMRS) started in July 2020. The university's robotics lab began by adapting water-resistant modules it had used in the past to allow the robot to operate in less than ideal conditions. They then added a series of turbines and thrusters so that it could move underwater.