Cloud Security Alliance Broadens Concentrate On Governance and Assurance for Agentic AI Systems
The Cloud Security Alliance (CSA) recently announced a series of CSAI Foundation milestones focused on securing what it calls the agentic control aircraft, including a brand-new catastrophic threat initiative, CVE Numbering Authority authorization, and the acquisition of 2 agentic AI requirements.
The April 29 statement, made at the CSA Agentic AI Security Top, centers on governance and assurance for agentic AI systems. CSA stated the milestones broaden the CSAI Structure’s 2026 mission of “Securing the Agentic Control Plane.”
According to CSA, the announcements consist of the launch of the STAR for AI Catastrophic Threat Annex, permission as a CVE Numbering Authority through MITRE and the acquisition of the Autonomous Action Runtime Management specification and Agentic Trust Framework.
“The global economy is competing with 2 exponentials at the same time: frontier models leapfrogging each other month over month, and viral, bottom-up adoption of agents inside the business,” said Jim Reavis, CEO and co-founder of CSA. “Today’s announcements provide enterprises, auditors, and regulators the technical specs and guarantee scaffolding to state yes to agentic AI without losing control of it.”
Catastrophic Danger Annex Planned
The STAR for AI Catastrophic Risk Annex is being released with support from Coefficient Offering, which CSA described as a humanitarian company support long-horizon AI security work. CSA stated the annex extends the AI Controls Matrix and STAR for AI assurance program to cover situations including loss of human oversight, uncontrolled system habits and other massive, permanent, society-wide consequences.
The annex is designed to focus on controls that can be tested in production environments, according to CSA. An associated CSA blog post said the task will identify existing AICM manages pertinent to catastrophic risk, introduce new controls where spaces exist, and specify proof requirements and screening requirements appropriate for independent evaluation.
The rollout is prepared in 4 stages from June 2026 through December 2027. Phase 1, from June through September 2026, is meant to translate disastrous risk scenarios into auditable control language. Stage 2, from October through December 2026, is intended to develop validation procedures. Stage 3, from January through June 2027, is planned to bring the annex into real-world environments through pilot evaluations, assessor training, and recommendation applications. Phase 4, from July through December 2027, is planned to produce public STAR for AI computer registry entries, benchmarking, and a State of Catastrophic AI Risk Controls Report.
CSA said the annex will line up with the NIST AI RMF, the EU AI Act and ISO/IEC 42001. The source does not record particular control text for the annex.
AICM and STAR for AI Context
The annex builds on CSA’s AI Controls Matrix, which CSA describes as a vendor-agnostic structure for cloud-based AI systems. CSA says the AICM includes 243 control objectives throughout 18 security domains and maps to standards including ISO 42001, ISO 27001, NIST AI RMF 1.0, and BSI AIC4.
The AICM bundle includes the matrix itself, mapping to NIST AI 600-1, ISO 42001, and the EU AI Act, execution standards, auditing standards, the AI-CAIQ survey, introductory guidance, and a STAR for AI Level 1 submission guide, according to CSA.