NSA Releases New Guidance for AI Data Security
The NSA’s Artificial Intelligence Security Center, in conjunction with other agencies, has released a new Cybersecurity Information Sheet (CSI) with recommendations for secure handling of data used in the development, testing, and operation of AI-based systems.
“AI Data Security: Best Practices for Securing Data Used to Train & Operate AI Systems” provides an overview of the AI system lifecycle and focuses on three particular areas of data security risk: data supply chain, maliciously modified (“poisoned”) data, and data drift.
Along with specific risks and mitigation strategies, the CSI outlines 10 broad steps for secure data handling:
- Verify that data sources use trusted, reliable, and accurate data.
- Verify and maintain data integrity during storage and transport.
- Use digital signatures to authenticate trusted data revisions.
- Leverage trusted infrastructure.
- Classify data and use access controls.
- Encrypt data.
- Store data securely.
- Leverage privacy-preserving techniques.
- Delete data securely.
- Conduct ongoing data security risk assessments.
This document is jointly authorized by the Cybersecurity and Infrastructure Agency (CISA), Federal Bureau of Investigation (FBI), Australian Signals Directorate’s Australian Cyber Security Center, New Zealand’s National Cyber Security Center, and UK’s National Cyber Security Center.
Learn more from the NSA.
Subscribe to our ADMIN Newsletters
Subscribe to our Linux Newsletters
Find Linux and Open Source Jobs
Most Popular
Support Our Work
ADMIN content is made possible with support from readers like you. Please consider contributing when you've found an article to be beneficial.
