Bolstering Cybersecurity: DISA's Plans to Integrate AI in Cybersecurity Efforts
According to reports by Patrick Tucker of Defense One, the Defense Information Systems Agency (DISA) is advancing its cybersecurity strategies by incorporating artificial intelligence (AI) for enhanced intrusion detection on Department of Defense (DOD) networks.
Deepak Seth, the technical lead for emerging technologies at DISA, highlighted at the AFCEA’s TechnetCyber conference that the agency intends to utilize AI models to expedite the analysis of vast data collected from various DOD network endpoints. This initiative is a part of their collaboration with DARPA on the Cyber Hunting at Scale, or CHASE, program.
Seth posed the question: "How can we use AI to process all this data?" He believes that AI technology can augment current procedures by detecting unknown threats much quicker than human operators.
However, DISA's aspirations go beyond merely automating anomaly detection. The agency aims to facilitate automated penetration testing on DOD networks, effectively simulating cyberattacks on itself to uncover vulnerabilities.
Eric Mellot, DISA’s senior technical strategist, clarified that the goal is to automate tasks traditionally carried out by a team of penetration testers. Mellot emphasized the growing need to implement "autonomous continuous validation," arguing that the incorporation of AI could bring about an AI model that "thinks like a hacker."
The Pentagon's previous endeavors underscored that red teaming, i.e., continuous attempts to breach Defense Department networks, improved cybersecurity more efficiently than merely conducting intermittent systems checks.
Seth also pointed out the relevance of recent AI advancements in the commercial sector, notably the ChatGPT platform from OpenAI, which exemplifies the rapidly evolving AI landscape. He believes these breakthroughs might ease the Pentagon's AI adoption by lowering the required skill level to experiment.
Notwithstanding, DISA is conscious of the potential threats posed by adversarial use of AI by nations such as China, coupled with emerging technologies like quantum computing. One of their key future research efforts involves the development of encryption forms resilient to AI models operating on advanced (though hypothetical) quantum computers.
Echoing this sentiment, Seth voiced his concerns over the possibility of quantum computers breaking RSA public key encryption protocols. The agency is collaborating with NIST, which recently introduced new algorithms designed to bolster data security in scenarios where traditional encryption measures fail.
Ultimately, DISA seeks to refine these algorithms for use within the Defense Department. Seth concluded that the agency is beginning to explore more secure means of key distribution for their optical backbone transport network and devices, signifying the early stages of DISA's appraisal of potential impacts.
As DISA continues its journey towards a more secure digital environment, the integration of AI in its cybersecurity strategies represents a significant leap forward. The aim is to not just stay abreast of technological advancements but to ensure the nation's defense systems are equipped to face the challenges of the future.