3 minute read
Artificial Intelligence is no longer a technology of the future. It’s here, shaping decisions in finance, healthcare, manufacturing, infrastructure and even the way we interact with knowledge every day. Yet as AI becomes more embedded in society, one question has become impossible to ignore: How do we know we can trust it?
To explore this challenge, we hosted a Scientific Meeting at Bushy House last year, bringing together experts from government, industry and academia, to examine how we build confidence in AI through robust measurement.
We have now published a report that captures the key insights and sets out a path forward. [link to report]
Despite rapid breakthroughs in generative AI and large language models the gap between cutting edge research and trustworthy technology is growing. Organisations want to adopt AI but lack clarity on:
how to evaluate it,
how to choose the right metrics,
how to ensure it’s safe, fair and reliable, and
where accountability should sit.
This report delves into these challenges and outlines five guiding principles designed to move the UK towards a robust AI assurance ecosystem:
Appropriate communication, education and accessible standards
Make AI concepts understandable and usable across sectors and professions.
Context dependent, risk based assurance
Evaluate AI based on its intended purpose and the potential consequences if it fails.
System level, outcome based metrics
Shift from measuring models to measuring real world impacts.
Transparency, explainability and accountability
Build mechanisms that support trust, oversight and responsible use.
Clear evaluation frameworks across the AI lifecycle
Assure AI not just before adoption, but continuously, with traceability and feedback loops.
These principles reflect not only the consensus of experts, but also the urgency of the moment. AI innovation is accelerating. Standards and governance must keep pace.
As the UK’s National Metrology Institute, we have a long history of building confidence in new technologies through clear measurement, robust standards and trusted testing frameworks. Just as metrology laid the foundations for safe aviation, pharmaceuticals, engineering and many other fields, it now has a central role to play in ensuring AI is safe and trustworthy. Our Getting the Measure of AI report provides analysis of the current assurance landscape, the challenges we face, and the opportunities ahead for the UK to lead in trustworthy AI.
16 Mar 2026