Global AI Assurance Pilot

Technical Testing of Generative AI Applications

The AI Assurance Pilot is a global initiative that aims to help codify emerging norms and best practices around technical testing of Generative AI applications. The Pilot ran from February to May 2025.

The Pilot Aims to:

Pair AI assurance and testing providers with organisations deploying Generative AI applications

Focus on technical testing of the real-life application (not the underlying foundation model)

Use the lessons learnt from specific examples to create generalisable insights on “what and how to test”

Participants

From different countries, representing diverse industries and use cases

Pilot Scope

Type of use cases

Risk dimensions to consider during testing (not exhaustive)

Safety and health risks
Unfair treatment of staff/ customers/ citizens
(Lack of) Transparency and recourse
Inappropriate data disclosure
Malicious use
Other security risks
Trust/ reputation concerns
Financial loss
(Lack of) Appropriate level of human oversight
Breach of other industry-specific (non-AI) regulatory requirements
Breach of other internal compliance requirements

Benefits

For AI assurance and testing vendors

For firms deploying Generative AI applications

Showcase of pilot at Asia Tech x Singapore

The insights and use cases from the pilot was published and showcased at Asia Tech x Singapore 2025.

Findings from the pilot were unveiled at Asia Tech x Singapore 2025, where real-world use cases and insights were shared with industry, government, and research communities. The showcase featured the collaborations between 17 pairs of AI deployers and their testing partners from diverse sectors, offering practical insights into how assurance practices can make GenAI more trustworthy and transparent.

Resources

Pilot
Insights

Case Study
Compendium

Pilot Process Overview

Pilot Participation Details