Nakasone (pictured) noted in his remarks that the US Intelligence Community, as well as the Department of Defense, have been using artificial intelligence for quite some time. Thus, artificial intelligence systems are already integral in managing and analyzing information on a daily basis. In doing so, such systems contribute in important ways to the decision-making by the NSA’s human personnel. At the same time, the NSA has been using artificial intelligence to develop and define best-practices guidelines and principles for intelligence methodologies and evaluation.

Currently, the United States maintains a clear advantage in artificial intelligence over is adversaries, Nakasone said. However, that advantage “should not be taken for granted”. As artificial intelligence organizational principles are increasingly integrated into the day-to-day functions of the intelligence and security enterprise, new risks are emerging by that very use. For this reason, the NSA has launched its new Artificial Intelligence Security Center within its existing Cybersecurity Collaboration Center. The mission of the Cybersecurity Collaboration Center is to develop links with the private sector in the US and its partner nations to “secure emerging technologies” and “harden the US Defense Industrial Base”.

Nakasone added that the decision to create the Artificial Intelligence Security Center resulted from an NSA study, which alerted officials to the national security challenges stemming from adversarial attacks against the artificial intelligence models that are currently in use. These attacks, focusing on sabotage or theft of critical artificial intelligence technologies, could originate from other generative artificial intelligence technologies that are under the command of adversarial actors.

Last Wednesday, the CIA’s Raman discussed some of the ways that artificial intelligence is currently being put to use by her agency to improve its analytical and operational capabilities. Raman noted that the CIA is developing an artificial intelligence chatbot, which is meant to help its analysts refine their research and analytical writing capabilities. Additionally, artificial intelligence systems are being used to analyze quantities of collected data that are too large for human analysts to manage. By devoting artificial intelligence resources to the relatively menial and low-level tasks of data-sifting and sorting, the CIA enables its analysts to dedicate more time to strategic-level products.

At the same time, however, the CIA is concerned about the rapid development of artificial intelligence by nations such as China and Russia, Raman said. New capabilities in artificial intelligence, especially the generative kind, will inevitably provide US adversaries with tools and capabilities that will challenge American national security in the coming years, she concluded.