Defense is one of the most controversial sectors in the global economy. Meanwhile, within the confines of tech, concepts of artificial intelligence are among the field's most compelling and debate-provoking matters, with some thinking advanced AI isn't far away and others expecting it not to be truly "conscious" for many years. Any context in which these two subjects find themselves together is bound to capture the attention of marketers, manufacturers, government officials and the general public.
The real question, then, is how exactly AI will affect defense. No matter how you look at it, the two have become notably intertwined. As governments are funneling billions or even trillions into their defense budgets and manufacturers are competing fiercely for the industry's available revenue wherever it may be, AI will be a major catalyst for a significant number of the sector's developments. Some of these technological advances have already manifested themselves in various forms. Others are currently more in the "potential" stage rather than being anywhere near commercial deployment. All of them, however, are worth paying attention to for businesses directly or indirectly connected to the defense industry.
U.S. commits significant funds to AI defense research
The U.S. undeniably holds lead position as the biggest national defense spender on the planet: $610 billion in 2017, according to the Stockholm International Peace Research Institute; more than the next seven highest-spending countries combined. It thus comes as little surprise that the Department of Defense recently increased its investment in AI-driven defense and reconnaissance projects, as The Washington Post reported in early September 2018.
According to the news provider, the DoD intends to add $2 billion in research and development spending to its existing AI budget over the course of the next five years. This comes directly at the behest of leaders at the agency's highest levels - as well as the White House, which has recently issued statements emphasizing the importance of AI throughout all aspects of American business and society. Additionally, AI initiatives originating from either the DoD or elsewhere in the military will now be overseen by a new agency, the Joint Artificial Intelligence Center. In a letter announcing the formation of this group, military leaders stated that AI had the capability to "change society, and, ultimately, the character of war."
Expanding on existing concepts
It's worth noting that AI is already in use within a number of well-known defense tools. Unmanned aerial vehicles, for example, have been a staple of defense technologies in the U.S. and elsewhere in the world for at least a decade. Forbes pointed out that in recent years, AI enhancements have allowed these drones to handle numerous aspects of their operation and navigation autonomously, including espionage, peacekeeping and various search tasks. As engineers and programmers develop new capabilities for the artificial neural networks and other core processes of AI, drones' defense-related functionality will likely broaden considerably.
The growing frequency of state-sponsored cyberattacks has also provided an opportunity for AI to serve as a security enhancement for web portals through which military data is transferred, as well as improved encryption and analysis of that information. This leads to more informed, sound decisions by officers on the ground, and the increased reliability of future AI-derived data may strip away a great deal of the uncertainty from certain military operations that might otherwise be unjustifiably dangerous.
Intriguing possibilities for the future
Tech experts are developing neural networks with greater complexity and scope even as we speak, according to a blog post by the RAND Corporation, with the goal of increasing their deep learning capabilities to a level where multi-layered data processing is an easy, even routine process. On a more long-term basis, everyone involved in the creation of AI technologies wants to achieve what is called artificial general intelligence: AI that can effectively think for itself and make decisions as a result of those conscious processes. AGI would, in theory, be able to pass the Turing test.
So what might this mean for the ground, naval and aerial defense branches of the world? For one, RAND stated that advanced AI could considerably simplify training operations and perhaps even cost less than "frozen" software platforms, which are obviously limited by the skill of their programmers and the hardware running them. But natural language processing, facial recognition and surveillance data are just a handful of the AI applications with major benefit for the military: Missiles ordered to launch with voice commands and guided by high-echelon deep learning, for example, could hit targets with extreme accuracy and potentially limit civilian casualties and collateral damage to nearby property. Further in the future, robots - currently used for soldier assistance, explosive disarmament and search-and-rescue operations - could perhaps even be used in combat, though this is perhaps decades away and will require considerable regulation.
Ethical issues
Questions of legal and moral culpability arise when considering the defense industry. For many AI-producing companies, including giants like Google's DeepMind and lesser-known but highly influential firms like Sentient Technologies, creating (or assisting) lethal weapons is a bridge they don't want to cross with their neural networks, and they signed a pledge with the nonprofit Future of Life Institute not to build such technologies.
According to Defense One, combat AI could make mistakes, such as hitting the wrong target, due to faults like imperfect imaging capabilities, but this will become less likely as neural networks improve. The truest key to AI's success in any operation, including defense, is the quality of data influencing its decisions. Advances in data mining and improved human analysis of such information will thus be pivotal.
