Invitation to Pilot AI Governance Testing Framework & Toolkit
INVITATION TO PILOT AI GOVERNANCE TESTING FRAMEWORK & TOOLKIT
This document is an invitation by Singapore's IMDA and PDPC to participate in a pilot phase for a Minimum Viable Product (MVP) AI Governance Testing Framework and Toolkit. The framework translates international AI ethics principles into testable criteria evaluated through a combination of technical tests and process checks. It provides tools for assessing fairness, explainability, and robustness, while fostering an AI testing community to shape future standards.
Summary and content fields are drawn from the document text and are pending editorial review.
Who issued it
Who it governs
AI system owners, developers, and technology solution providers interested in piloting and validating AI governance testing tools.
What the document contains
The document covers the objectives, structure, and pilot invitation for an AI Governance Testing Framework and Toolkit Minimum Viable Product. It details eight core AI ethics principles, dividing them into technical tests and process checks, and outlines the included open-source testing libraries and limitations.
The MVP does not define ethical standards, does not guarantee that tested AI systems are completely free from risks or biases, and does not provide numerical thresholds for metrics.
- Singapore's IMDA and PDPC have developed an AI Governance Testing Framework and Toolkit MVP to enable objective testing of responsible AI practices.
- The MVP incorporates 8 selected AI ethics principles evaluated through a mix of technical tests and process checks.
- The toolkit integrates popular open-source libraries such as SHAP, AIF360, Fairlearn, and the Adversarial Robustness Toolbox.
- Participating in the pilot phase allows organizations to conduct self-testing, build stakeholder trust, and provide feedback to shape future versions.
- The initiative aims to foster an AI testing community to help develop industry benchmarks and contribute to international standards.
· PART I: WHY DO WE NEED TESTING FOR AI GOVERNANCE? · PART II: WHAT IS THE AI GOVERNANCE TESTING FRAMEWORK AND TOOLKIT? · PART III: INVITATION TO PILOT THE MVP
How it scores
Each score is backed by a quote verified against the document text. Editorial review has not yet confirmed them.
Accountability
4/5Clear responsibility for AI outcomes
Accountability – Assessed through process checks of documentary evidence, including evidence of clear internal governance mechanisms for proper management oversight of the AI system’s development and deployment.
Paragraph 13(d)
screening score · 2026-10-23
Human Dignity
0/5Respect for the inherent worth of persons
screening score · 2026-10-23
Human Oversight
4/5Humans remain in control of consequential decisions
This includes defining role of human in its oversight and control of the AI system such as human-in-the-loop, human-over-the-loop, or human-out-of-the-loop.
Paragraph 13(e)
screening score · 2026-10-23
Privacy
2/5Protection of personal data and private life
Leverage existing testing and certification regimes and efforts, i.e., cybersecurity and data governance (including data protection/privacy) and not reinventing the wheel
Paragraph 10(c)
screening score · 2026-10-23
Enforcement
1/5Consequences and compliance mechanisms with real force
Voluntary AI governance frameworks and guidelines have been published to help system owners and developers implement trustworthy AI products and services.
Paragraph 2
screening score · 2026-10-23
Impact Assessment
4/5Required assessment of an AI system's risks/impacts before or during use
Process checks include verifying documentary evidence of considerations given to the choice of models, such as rationale, risk assessments, and trade-offs of the AI model.
Paragraph 20
screening score · 2026-10-23