[ad_1]
At the moment, in a landmark collaboration, the U.S. Cybersecurity and Infrastructure Safety Company (CISA) and the UK Nationwide Cyber Safety Centre (NCSC) are proud to announce the discharge of the Guidelines for Secure AI System Development. Co-sealed by 23 home and worldwide cybersecurity organizations, this publication marks a major step in addressing the intersection of synthetic intelligence (AI), cybersecurity, and important infrastructure.
The Pointers, complementing the U.S. Voluntary Commitments on Ensuring Safe, Secure, and Trustworthy AI, present important suggestions for AI system improvement and emphasize the significance of adhering to Secure by Design ideas. The strategy prioritizes possession of safety outcomes for purchasers, embraces radical transparency and accountability, and establishes organizational buildings the place safe design is a prime precedence.
The Pointers apply to all sorts of AI methods, not simply frontier fashions. We offer options and mitigations that can assist knowledge scientists, builders, managers, decision-makers, and danger homeowners make knowledgeable selections concerning the safe design, mannequin improvement, system improvement, deployment, and operation of their machine studying AI methods.
This doc is aimed primarily at suppliers of AI methods, whether or not primarily based on fashions hosted by a corporation or making use of exterior software programming interfaces. Nevertheless, we urge all stakeholders—together with knowledge scientists, builders, managers, decision-makers, and danger homeowners make—to learn this steerage to assist them make knowledgeable selections concerning the design, deployment, and operation of their machine studying AI methods.
CISA invitations stakeholders, companions, and the general public to discover the Guidelines for Secure AI System Development in addition to our not too long ago revealed Roadmap for AI to study extra about our strategic imaginative and prescient for AI know-how and cybersecurity. To entry study extra, go to CISA.gov/AI.
[ad_2]
Source link