The U.S. National Institute of Standards and Technology (NIST) has started a new program to address the growing role of artificial intelligence (AI) in cybersecurity and privacy. The program is run through the National Cybersecurity Center of Excellence (NCCoE). It aims to deal with the “cybersecurity of AI and AI for cybersecurity.” It will guide the use of NIST’s Cybersecurity Framework (CSF) 2.0.
The program is a response to the fast integration of AI into various areas of business and society.
This includes using AI in cybersecurity technologies, using generative AI tools like ChatGPT in workplaces, and its use by cybercriminals for phishing and making deepfakes. Katerina Megas, program manager for NIST’s AI and cybersecurity programs, stressed the urgent need for standards, guidelines, tools, and best practices to manage cybersecurity and privacy in the AI era. She said the program aims to ensure the responsible use of AI for cybersecurity and privacy protection.
It also identifies key actions organizations must take to counter AI-enabled threats. The program addresses various risks and benefits of AI, such as:
– The re-identification of private information from AI training datasets
– AI’s role in behavioral analysis and surveillance
– Its use in cybersecurity tasks like threat hunting
– Its help to users navigating privacy preferences online
A key feature of NIST’s CSF 2.0 is the introduction of community profiles. These tailor the framework to specific business sectors, threat types, and technology uses.
NIST’s AI cybersecurity initiative launched
The new AI profile will focus on:
1. Addressing cybersecurity and privacy risks at organizations by securing AI components and minimizing data leakage.
2. Defending against AI-enabled cyberattacks. 3.
Enhancing cyber defense activities and privacy protections through AI-powered assistance. Developing this community profile is meant to adapt the CSF 2.0 to AI use contexts. It will eventually help integrate other frameworks with different AI tools.
This program builds on NIST’s previous work related to AI safety and privacy. In January 2024, NIST published “Adversarial Machine Learning: A Taxonomy and Terminology of Attacks and Mitigations.” In July 2024, it partnered with generative AI companies OpenAI and Anthropic to evaluate their new models.
