Call for Evidence: AI testing, evaluation and assurance in the UK

In September 2025, the Department for Science, Innovation and Technology (DSIT) published its Trusted Third-Party AI Assurance Roadmap, setting out the Government's ambition to grow a credible, high-quality third-party AI assurance market. The roadmap identifies several barriers to that market, including:

• The absence of an agreed basis for judging the quality and comparability of AI assurance goods and services, and the technical standards that should underpin them;
• Shortages of skills and talent;
• Gaps in assurance providers' access to the information they need about the AI systems they assess; and
• Limited support for developing innovative testing and evaluation methods as AI capabilities advance.

Through its applied assurance and industry support work, the Centre for AI Measurement aims to address the technical dimensions of these challenges – in particular, the methods, tools and evidence needed for developing and adopting rigorous and comparable testing and evaluation techniques. 

This Call for Evidence gathers input from across the ecosystem so that this work reflects the practical needs and constraints of those developing, deploying and assuring AI systems. Your responses will inform the design of the Centre's industry support work, including the problems it addresses first, the sectors it prioritises and the form that support takes. For organisations interested in engaging further, the Call for Evidence also provides an opportunity to signal interest in future Centre activities and stay informed as the programme develops and call for applications open.

For the purposes of this call for evidence, AI assurance refers to the process of measuring, evaluating and communicating the trustworthiness of AI systems and their components.

Assurance can be carried out at different points of the AI lifecycle and by different actors. For example, it may be conducted internally by development teams, by a buyer assessing a system during procurement, or independently by a third-party assurance provider as part of a formal audit or certification process. Independent, third-party assurance is central to building justified trust in AI and to growing a credible assurance market and is a priority for the UK Government.

This call is concerned primarily with the technical foundations that underpin these processes: the measurement science, methods, metrics, tools and infrastructure used to test and evaluate AI. The Centre's distinctive role is in measurement science - defining how AI should be measured and evaluated in ways that are scientifically rigorous, reproducible and comparable. Better measurement strengthens assurance wherever it is done, whether inside an organisation or by a third-party provider.

We are less focused, in this call, on the broader governance, legal, organisational and regulatory aspects of AI assurance. These are important elements in the AI assurance toolbox, but they are not the primary concern of this programme. Other initiatives address them, which we do not want to duplicate.