Nuclear AI is difficult, not impossible, new innovations prove
Lowther, PhD, and McGiffin. 2X MA, 19
(Adam, Director of Research and Education at the Louisiana Tech Research Institute, Curtis, School of Strategic Force Studies, at the Air Force Institute of Technology and an adjunct professor for Missouri State University’s Department of Defense and Strategic Studies where he teaches strategic nuclear deterrence theory and NC3 education https://warontherocks.com/2019/08/america-needs-a-dead-hand/)
There is a fourth option. The United States could develop an NC3 system based on artificial intelligence. Such an approach could overcome the attack-time compression challenge. DARPA’s Knowledge-directed Artificial Intelligence Reasoning Over Schemas program is an example of how an American NC3 system based on artificial intelligence might function. Fusing the contextual and temporal events of a nuclear attack into an analytic-based artificial intelligence capability may ensure rapid comprehension and in turn generate associated and prompt actionable responses. The biggest challenge for such a system is its ability to learn and adapt. Unlike the game of Go, which the current world champion is a supercomputer, Alpha Go Zero, that learned through an iterative process, in nuclear conflict there is no iterative learning process. Thus, a fully empowered “general” artificial intelligence system that learns may be far more difficult to design than a “narrow” artificial intelligence system that engages in limited analysis and decision-making. Artificial intelligence is perhaps best poised to assist humans when it comes to the dimensions of detecting a nuclear attack and deciding which planned option best meets the criteria designed by programmers. Here, artificial intelligence may, to a small degree, mitigate the tyranny of attack-time compression and accelerate wartime decision-making. However, when a president may have, at most, six minutes to make a decision, time compression still poses a fundamental problem. Artificial intelligence is already being used for target identification, controlling autonomous platforms, pattern recognition, and a number of other wartime tasks. It is capable of processing vast amounts of information very quickly and assessing the pros and cons of alternative actions in a thoroughly unemotional manner. According to Vincent Boulanin: Recent advances in artificial intelligence could be leveraged in all aspects of the nuclear enterprise. Machine learning could boost the detection capabilities of extant early warning systems and improve the possibility for human analysts to do a cross-analysis of intelligence, surveillance, and reconnaissance (ISR) data. Machine learning could be used to enhance the protection of the command and control architecture against cyberattacks and improve the way resources, including human forces, are managed. Machine learning advances could boost the capabilities of non-nuclear means of deterrence: be it conventional (air defence systems), electronic (jamming) or cyber. However, artificial intelligence is no panacea. Its failures are numerous. And the fact that there is profound concern by well-respected experts in the field that science fiction may become reality, because artificial intelligence designers cannot control their creation, should not be dismissed. For the United States, every option presents significant risk and uncertainty. Reality, however, is progressing to a point where the United States must address the challenge we outlined above. Russia and China are not constrained by the same moral dilemmas that keep Americans awake at night. Rather, they are focused on creating strategic advantage for their countries.