- Posted May 23, 2025
- Technical Advisories
Best Practices for Securing Data Used to Train & Operate AI Systems
For the opportunities of artificial intelligence to be fully realised, it must be developed and used in secure and responsible ways.
Data security is paramount in the development and deployment of AI systems.
The National Cyber Security Centre (NCSC) has co-authored cyber security information and guidance with the National Security Agency (NSA) and other international partners regarding securing data used in artificial intelligence (AI) and machine learning (ML) systems.
This guidance provides a brief overview of the AI system lifecycle and general best practices to secure data used during the development, testing, and operation of AI-based systems.
This information is intended primarily for organisations using AI systems in their operations, with a focus on protecting sensitive, proprietary, or mission-critical data.
There are principles outlined in this guidance that provide a robust foundation for securing AI data and ensuring the reliability and accuracy of AI-driven outcomes.
As organisations in various sectors rely more and more on AI-driven outcomes, ensuring data security becomes crucial for maintaining accuracy, reliability, and integrity.
Read the guidance: AI and Data Security.(external link)
For more NCSC NZ updates, follow us on LinkedIn.(external link)(external link)