Skip to content
The National Physical Laboratory The National Physical Laboratory
Accessibility and language tools
Data Science and AI

Centre for AI Measurement

Building confidence in AI through robust measurement

 

About the Centre for AI Measurement

The Centre for AI Measurement is a national initiative led by the National Physical Laboratory (NPL), developing the measurement science needed to test, evaluate and assure the quality and trustworthiness of AI systems. It forms part of the UK Government’s wider programme to support safe and responsible AI adoption and strengthen the emerging market for third-party AI assurance. 

Drawing on NPL’s role as the UK’s National Metrology Institute, the Centre will bring together collaborative research, practical industry application, and ecosystem coordination to advance rigorous, repeatable and comparable approaches to AI testing and evaluation and support their adoption in practice.

Robust, trusted tools for assessing the safety and effectiveness of emerging AI systems are essential for businesses to deploy them with confidence. The centre unites industry, start-ups and researchers to create new ways for organisations to identify and mitigate risks early.

Kanishka Narayan MP - Minister for AI

 
Our vision and mission
Our Vision

Confidence in AI is grounded in robust scientific evidence.

A future where decisions about AI are informed by reliable, comparable and scientifically robust evidence. Strong, impartial measurement foundations will support the responsible development, deployment and assurance of AI, enabling organisations to better understand how systems perform and behave in practice.

Our Mission

Build the scientific foundations for rigorous and scalable AI assurance.

We collaborate to develop and pilot robust measurement methods, validate them in real-world settings, and translate measurement science into practical guidance, technical specifications and international standards.

Why measurement matters

As AI systems become more capable and widely adopted, confidence in their safety, reliability and performance must be grounded in robust evidence. While existing governance frameworks and standards help define what trustworthy AI should achieve, a significant gap remains in how these properties are measured, evaluated and demonstrated in practice.

Measurement science provides the foundations for closing this gap. By establishing clear definitions of what is being measured, repeatable evaluation methods, and rigorous approaches to uncertainty and traceability, it helps ensure that AI testing results are reliable, reproducible and comparable. These foundations are essential for building a credible and scalable AI assurance ecosystem, where technical claims can be independently verified and evidence can be trusted across organisations and sectors.

Through the Centre for AI Measurement, NPL aims to underpin the UK's AI assurance ecosystem with robust measurement science. Supporting the delivery of the UK government's Trusted third-party AI assurance roadmap, a key priority is to drive the growth of the third-party AI assurance market by developing shared measurement methods, testing protocols and good practice. This will help independent assurance providers generate credible and comparable evidence, building confidence across the AI supply chain and enabling organisations to develop, procure and deploy AI with greater confidence.

 

Areas of work

1. Measurement science and research

Building the scientific foundations for robust AI testing, evaluation and monitoring, and translating them into good practice guidance and pre-standardisation work.

Read our report

2. Applied assurance and industry support

Working directly with industry, the public sector, and assurance providers to test assurance approaches across various sectors and use cases and support their adoption.

More information coming soon

3. Community building and coordination

Strengthening collaboration, knowledge sharing and capability building to support a more connected, capable, and growing UK AI assurance ecosystem.

Join our community

 
About NPL

As the UK’s National Metrology Institute, NPL provides the measurement capability that underpins confidence in technologies, products, services and data. For more than 125 years, it has helped translate scientific knowledge into practical measurement infrastructure supporting innovation, regulation, international trade and industrial competitiveness.

NPL’s impartiality, long-term national role and relationships across government, industry, academia, standards organisations and the international metrology community provide a strong basis for the Centre. As a neutral convenor, it can bring together the expertise needed to ensure that testing, evaluation and monitoring approaches are scientifically robust, responsive to real-world requirements and trusted by multiple stakeholders. Its enduring capability will help establish sustainable foundations for AI assurance that can evolve alongside technology over the coming decades.

 

From Frameworks to Protocols: A Measurement Science Foundation for Technical AI Assurance

Bridging the gap between AI governance frameworks and real-world assurance through the generation of repeatable and transparent evidence.

Read the report

AI Assurance Innovation Network

Be part of a growing community working to advance the science and practice of AI testing, evaluation and assurance.

Convened by the Centre for AI Measurement at NPL, the Network brings together industry, researchers, assurance providers, policymakers, regulators and standards bodies to share knowledge, exchange insights and advance practical approaches to AI assurance.

Connect with researchers, industry, government and assurance providers to tackle shared challenges in AI assurance.

Contribute your expertise to shape the Centre’s research priorities and inform pre-standardisation guidance and testing protocols.

Stay up to date with emerging research and good practice in AI Measurement and evaluation.

Join the AI Assurance Innovation Network

Sign up to receive updates, hear about upcoming activities and opportunities, and contribute to the work of the Centre for AI Measurement.


By signing up to the AI Assurance Innovation Network, you agree to receive email updates from the National Physical Laboratory (NPL) about the Centre for AI Measurement and Network activities, including news, publications, events and opportunities to get involved. To find out how we use and manage your data, please read our Privacy notice.