CISA Joins ACSC-led Guidance on How to Use AI Systems Securely

Summary:

The Australian Cyber Security Centre (ACSC) has published joint guidance with the Communications Security Establishment (CISA) on securely using AI systems. The guidance, collaboratively developed with international organizations, provides AI system users with an overview of AI-related threats and steps to manage AI-related risks. The guidance covers six AI-related threats, including data poisoning, input manipulation, and re-identification of anonymized data. The document aims to assist users of AI systems in understanding and mitigating risks, while encouraging developers to review Guidelines for Secure AI System Development. For more information on CISA's efforts to address AI cybersecurity opportunities and risks, visit CISA.gov/AI.