Singapore's AI Verify Testing Framework

Applies To

Overview

AI Verify is Singapore's AI governance testing framework, launched by the Infocomm Media Development Authority (IMDA) and the PDPC on 25 May 2022 as a minimum viable product, then updated and piloted internationally. The framework gives organisations a structured method to test and document AI systems against internationally recognised AI ethics principles, including transparency, explainability, fairness, safety, and accountability. AI Verify has two parts: automated technical tests of AI models, run through a software toolkit, and process checks of how AI development and deployment are governed and documented. Results are compiled into an AI Verify report, which organisations can share with regulators, business partners, or customers as evidence of responsible AI practices. The framework aligns with international references including the Organisation for Economic Co-operation and Development (OECD) AI Principles, the EU's ALTAI (Assessment List for Trustworthy Artificial Intelligence) guidelines, and ISO/IEC international standards. IMDA has worked with international partners on interoperability, including the US National Institute of Standards and Technology's AI Risk Management Framework (NIST AI RMF) and the UK Information Commissioner's Office (ICO). AI Verify is voluntary and does not carry enforcement consequences, but it is increasingly referenced in public sector procurement and financial sector guidance in Singapore. For large language models, IMDA and the AI Verify Foundation launched Project Moonshot, an evaluation toolkit, in open beta on 31 May 2024.

Key Requirements

• •Voluntary participation; no legally mandated obligation for private sector organisations
• •Organisations using the framework must test AI systems against eleven AI ethics principles including fairness, transparency, explainability, robustness, and data governance
• •Technical tests must be conducted using the AI Verify software toolkit or an equivalent approved methodology
• •Process checks require documented evidence of internal governance practices including risk management, human oversight, and stakeholder engagement
• •AI Verify reports must accurately reflect test results and must not be used in a misleading manner
• •Project Moonshot, launched in 2024, provides separate evaluation and red-teaming tools for large language models
• •Organisations should maintain records of testing artefacts sufficient to support third-party verification if requested

What Your Organization Must Do

• →Decide whether AI Verify testing adds value, since participation is voluntary and carries no enforcement consequences.
• →Download the AI Verify toolkit and run technical tests on priority AI models.
• →Complete the process checks by documenting governance, risk management, human oversight and stakeholder engagement.
• →Map existing controls to the eleven AI ethics principles, including fairness, transparency and data governance.
• →Review AI Verify reports before sharing, so results stay accurate and are not presented misleadingly.
• →Watch Singapore public sector procurement and financial sector guidance, which increasingly reference the framework.

Playbook Guidance

Step-by-step implementation guidance for compliance teams.