NSA Research Director Discusses the Impact of ChatGPT and AI on National Security
In late 2021, amidst the rise of ChatGPT and other groundbreaking AI products, Gilbert Herrera, an electrical engineer, was appointed as the research director of the US National Security Agency (NSA). In a recent interview, Herrera shared his insights on the significance of the “ChatGPT moment” and its implications for the NSA’s mission. The conversation has been slightly edited for clarity and brevity.
The Emergence of AI and Its Significance
Herrera emphasized that the NSA has long recognized the potential of artificial intelligence in enhancing its operations, making them smarter, faster, and more scalable. The agency has been actively involved in AI research for over two decades, contributing to the advancements that led to the current ”ChatGPT moment.”
While large language models have existed before the advent of generative pretrained (GPT) models, the ability to engage in conversational interactions and generate creative content, such as jokes, sets ChatGPT apart from previous AI systems.
Leveraging Commercially Available Language Models
Herrera acknowledged that commercially available large language models could be valuable to the NSA in various ways. One notable application is in the field of cybersecurity, where these models have demonstrated proficiency in identifying vulnerabilities in software code. However, he also cautioned about the potential risks associated with these models, such as the possibility of stealing or manipulating them through inversion attacks.
The first line of defense in AI security is good cybersecurity. It means protecting your models, protecting the data that’s in there, protecting them from being stolen or manipulated.
To mitigate these risks, Herrera emphasized the importance of robust cybersecurity measures, including safeguarding the models, the data they contain, and preventing unauthorized access or manipulation.
The NSA’s Role in Pioneering Technologies
Herrera alluded to the NSA’s history of developing cutting-edge technologies before they become widely known, such as public-key cryptography. However, he refrained from divulging specific details about the agency’s involvement in the development of AI technologies, citing the need for secrecy in matters related to national security.
As the NSA continues to navigate the rapidly evolving landscape of artificial intelligence, Herrera’s insights shed light on the agency’s approach to leveraging these technologies while addressing the associated challenges and risks. The “ChatGPT moment” marks a significant milestone in the advancement of AI, and its impact on national security will undoubtedly shape the future of the NSA’s operations.
2 Comments
So, the NSA thinks private data is the new gold rush for AI supremacy, huh? Interesting.
Well, looks like our personal info is now the secret sauce to AI dominance, watch out!