Bushy House Scientific
meetings

Fostering collaboration between far-reaching scientific areas

The National Physical Laboratory prides itself on enabling innovation through creating discourse and fostering collaboration between experts from diverse scientific areas. As such, we are pleased to launch our Scientific Meeting series, reviving the historic NPL symposium series which ran between 1953-1963. Bringing together multidisciplinary experts on topics of national or international importance where ground-breaking advances, more demanding requirements or a rapidly changing scientific landscape requires in depth discussion, the series is designed to create defining moments for the scientific community at large.

Getting the measure of AI: trustworthy and impactful AI through metrics and assurance

21-22 October – Bushy House, Teddington  

Artificial Intelligence (AI) is a transformative technology that is becoming increasingly integrated into daily operations and decision-making processes. As AI touches more aspects of our daily lives, particularly safety-critical and highly regulated areas, it is paramount that we have confidence in the effective, trustworthy and sustainable use of AI. 

Governments across the globe are seeking ways to deliver this confidence and assurance in algorithm outputs. The UK is leveraging the expertise of its existing certification bodies for specific application areas, whereas the EU has introduced a broader AI act that mandates risk management and conformity assessment. 

These approaches require robust, scientifically proven measurement tools and global technical standards – which bring more questions than answers:

• How do we measure trustworthiness of AI or the quality of training data? 
• What measurement standards do we need, and can we, develop? 
• Can we be confident in the interpretation of AI outputs by a non-expert user? 
• What types of metrics and benchmarks are needed in different sectors and for different use cases? 
• What is needed to foster development, acceptance and uptake of robust AI assurance tools and metrics? 

At heart, many of these questions are metrological (the science of measurement) in nature. Answering them will need collaboration across end users, regulators, government, developers, and metrologists, each bringing their own expertise to the table. This meeting aims to provide a place for those discussions to start. 

This two-day discussion will explore the question of metrology for AI, specifically looking at the role of metrology, metrics, and benchmarks within the AI assurance process. Convening influential stakeholders from across the AI governance, innovation and adoption chain (including government, industry, and regulators), we will delve into the measurement science behind trustworthy AI. We are looking to set the direction, foster partnerships and collaboration, and build knowledge across sectors on how the UK can be a global leader in AI assurance and underlying standards.

NPL Scientific meetings are invite-only events.

Be part of the Bushy House scientific meetings

Submit an idea or topic for a Scientific Meeting