During the AI Safety Summit, which took place in early November 2023 in the UK and was dedicated to the issues of AI, its development and security, the importance of cyber security and the "safe design" approach to the development of artificial intelligence were separately emphasized.
Therefore, the National Cyber Security Center of Great Britain (NCSC) together with the Agency for Cyber Security and Infrastructure Security of the USA (CISA) presented the first document of its kind - Recommendations for the development of secure artificial intelligence systems . Leading industry experts, 21 international agencies and ministries from around the world participated in the development of the Recommendations - including all members of the G7 group of countries and the Global South. So far, 18 countries have approved the Recommendations.
" These guidelines mark a significant step in building a truly global common understanding of AI-related cyber risks and mitigation strategies to ensure that security is not an afterthought but a core requirement throughout, " said Lindy Cameron , NCSC CEO. .
The guidelines provide behaviors to help improve security and cover four key areas within the AI system development lifecycle, namely:
- Safe design.
The chapter provides guidance for the design phase of AI systems and covers risk understanding and threat modeling, as well as specific topics and tradeoffs to consider in system and model design.
- Safe development.
The chapter covers the development phase of an AI system, including supply chain security, documentation, and technical management
- Secure deployment.
This section provides guidance that applies to the deployment phase of AI system development, including protecting infrastructure and models from compromise, threat, or loss, developing incident management processes, and responsible release.
- Safe operation and maintenance.
The section is dedicated to the guidelines that apply to the safe operation and maintenance stage of the AI system development cycle. It provides guidance on activities particularly relevant after system deployment, including logging and monitoring, update management, and information sharing.
The guidelines are the first of their kind to be globally agreed upon. They will help developers of AI systems make informed cybersecurity decisions at every stage of the development process – whether those systems are built from scratch or built on tools and services provided by others.
"This joint effort reaffirms our mission to protect critical infrastructure and reinforces the importance of international partnerships to secure our digital future," summarizes Jen Easterly, Director of CISA.
Recall that on November 1 and 2, 2023, the UK hosted the first AI Security Summit, which brought together governments of many countries, leading technology organizations, academia and civil society to inform rapid national and international action at the forefront of AI development . As a result of the meeting, the Bletchley Declaration was adopted , which was signed by 28 countries, including Ukraine.
More: https://entc.com.ua/uk/2053-yuryst-konsultuye-rekomendatsii-shchodo-bezpeky-shi
