CSA STAR for AI Certifications: What are They?
Published 11/17/2026
The CSA STAR for AI program extends STAR, the largest cloud assurance program in the world, to artificial intelligence. Any organization that develops, orchestrates, or delivers AI services can benefit from completing a STAR for AI certification. These certifications are based on the AI Controls Matrix (AICM), which preserves all controls of the Cloud Controls Matrix (CCM) and adds AI-specific controls covering risks such as adversarial inputs, data poisoning, model integrity, explainability, human oversight, and AI supply chain transparency.
In this blog, learn more about STAR for AI certifications and what’s required to complete them.
What are the certifications under the STAR for AI program?
STAR for AI follows the same structure as STAR. Its certifications are designed to help AI service providers demonstrate how they secure and govern their AI services, and to give customers, regulators, and partners a trusted way to verify it. They are divided into two levels:
- Level 1, self-assessment
- Level 2, independent third-party certification or attestation
What is STAR for AI Level 1?
At Level 1, organizations evaluate and document the AI security and governance controls that apply to them using the AI Consensus Assessment Initiative Questionnaire (AI-CAIQ), which translates the AICM controls into a standardized set of questions. Completed AI-CAIQs are submitted to the STAR Registry, where they become publicly available, giving customers visibility into the provider’s AI practices.
Organizations can optionally run their AI-CAIQ through Valid-AI-ted, CSA’s AI-based scoring service, which evaluates the responses and returns a score and structured feedback before or alongside submission.
Organizations should pursue Level 1 if they are:
- Operating in a lower-risk environment, or at an early stage of their AI governance program
- Wanting to offer transparency around the AI controls they have in place
- Looking for a cost-effective first step toward Level 2
What is STAR for AI Level 2?
STAR for AI Level 2 consists of two independent third-party assessments: STAR for AI Certification and STAR for AI Attestation. There are associated fees for Level 2.
STAR for AI Certification is performed by an accredited certification body and combines a management system standard with the AICM, through one of three paths:
- ISO/IEC 27001 and AICM
- ISO/IEC 42001 and AICM
- ISO/IEC 27001, ISO/IEC 42001, and AICM
Certificates follow the normal protocol of the underlying ISO standard.
STAR for AI Attestation is performed by a licensed CPA firm as a SOC 2 engagement that uses criteria from the AICPA together with the AICM.
Organizations should pursue Level 2 if they are:
- Operating in a medium to high-risk environment, or serving customers who require independent assurance
- Already holding or adhering to ISO/IEC 27001, ISO/IEC 42001, or SOC 2
- Looking for a cost-effective way to extend existing assurance to their AI services
What are the scopes of the certificates?
Both STAR for AI Level 1 and Level 2 focus on AI services. A listing reflects the assessed scope, not the organization as a whole. A typical scope should include:
- The AI services or products included
- The models used to deliver them, including third-party models
- The data pipelines that feed those models
- The infrastructure on which the models run
- The departments, organizational units, and locations included
- Exclusions, with their justification
What are the prerequisites for obtaining a certificate?
For STAR for AI Level 1, there are no prerequisites. For Level 2, the prerequisite is a STAR for AI Level 1 listing, with an AI-CAIQ evaluated through Valid-AI-ted.
What types of organizations receive STAR for AI certificates?
STAR for AI is intended for organizations that perform one or more AI service roles defined in the AICM:
- Model Providers, which develop and deliver AI models
- Orchestrated Service Providers, which deliver AI platforms, orchestration layers, or managed AI services
- Application Providers, which deliver end-user AI applications, including AI-powered SaaS
A single organization often performs several of these roles at once, for example when it self-hosts an open-weight model and builds an application on top of it.
Who issues the certificates?
STAR for AI assessments are conducted by CSA-approved assessment firms:
- Certification bodies must hold ISO/IEC 17021-1-based accreditation covering ISO/IEC 27001 (per ISO/IEC 27006-1) and/or ISO/IEC 42001 (per ISO/IEC 42006). Their audit teams must hold the CCSK and STAR AI Controls Auditor training certificates.
- CPA firms performing attestations must ensure that the engagement team lead and partner hold the CCSK and STAR AI Controls Auditor training certificates.
Requirements and guidance for auditors and organizations are available from CSA, including:
- AICM Auditing Guidelines
- STAR Auditor Training, including STAR AI Controls Auditing
- STAR and STAR for AI Valid-AI-ted Guide
- STAR to STAR for AI Transition Guidance
What are the certificate validity periods?
STAR for AI self-assessments (Level 1) are valid for one year. STAR for AI Certifications (Level 2) are valid for three years. STAR for AI Attestations (Level 2) are valid for one year, at which time a complete re-evaluation is performed.
Learn more about the STAR for AI program and certifications here.
Unlock Cloud Security Insights
Subscribe to our newsletter for the latest expert trends and updates
Related Articles:
AI Governance Is Cloud Governance: From Posture to Proof
Published: 11/17/2026
AI Doesn’t Just Need Guardrails. It Actually Needs Independent Assurance.
Published: 11/16/2026
Cloud Security Alliance Announces Winners of 2026 Pitchapalooza
Published: 11/12/2026
.jpg)


.jpg)

.png)
.jpeg)
.jpeg)
.jpeg)
