Blog

What role does Machine Learning have in future Wars?

Jan, 2021 - By WMR

Ever since Arthur Samuel, the pioneer of computer gaming and artificial intelligence, popularized the term “machine leering”, we have explored the possibility of using it in almost every industry. Machine learning, for layman, refers to a type of artificial intelligence that enables software applications to become more precise by predicting outcomes without being explicitly programmed to do so. The primary objective of introducing machine learning is to optimize human capabilities and incorporate enhanced accuracy. For so much talk about these technologies, defense and military are one of the few sectors where a machine is not fully incorporated. That being said, researchers are now exploring the possible application of ML on battlefield.

One of the few things that need to be considered whether an algorithm should make life and death choices or humans do. There is still debate going on since when it comes to humans, there is always a question to maintain moral standing. For instance, if we consider a pilot chasing surface-to-air missiles, he needs to follow a certain group of protocols or steps before attacking the missiles. The pilot may or may not process the decision based on those protocols although he will take the decision with expected cost and benefits. There is a good chance of human mistake, lack of judgment, or simply bad luck where things can go terribly wrong. This where machine learning comes into play. An AI-enabled system will provide a list of alternatives, defining relevant possible results and conditional probability estimates with human inputs, if deemed necessary.

The U.S. Army is already taking a leap forward in terms of incorporating AI in the military. The SRI International has been working with the Army Research Laboratory as a part of the Internet of Battlefield of Things Collaborative Research Alliance. The program is focused on investigating methods to harden the machine learning algorithm to be less susceptible. Another goal is to create the next generation of algorithms that is resilient and robust. Researchers have found that machine learning better than humans when it comes to understanding the chain of events to undertake the best possible decision. For instance, human judgment mainly relies on the available examples in the memory instead of using more precise indicators such as historical data.

In such cases, the algorithm is far superior when it comes to understanding probabilities for a chain of events. As a matter of fact, a relatively simple combination of two conditional possibilities is significantly difficult for some individuals. At this time, it seems far more complicated and challenging. However, it is necessary that we incorporate these advanced technologies in the military. Russia and China have already started using machine learning to boost decision-making. The approach for this should be holistic and for which the military should require technology developers on different issues regarding the battlefield such as accountability and ethics. Developers should be able to demonstrate that the results generated with ML are not biased that would favor destruction. One way of doing this is by making machine learning-enabled systems for decision-making model that would require human accountability. That could truly be golden for the inculcation of this technology.