In this episode, Nicolas Cubaud, founder and CEO of EyeSnap and co-founder of TotalYmage, editor of the preliminary European norm (prEN) 18281 on artificial intelligence, evaluation methods for accurate computer vision systems. With over 25 years of experience in computer vision and image processing, Nicolas discusses how iSnap turns images and videos into actionable business insights for industrial and regulated sectors, such as nuclear conformity assessments and workplace safety monitoring (e.g., detecting proper PPE in hazardous areas).

The conversation explores the development of prEN 18281 (evaluation methods) and its companion prEN 18288 (taxonomy of computer vision tasks), covering classification, detection, segmentation, tracking, pose estimation, image/video generation, and metrics like IoU. Nicolas explains how these standards translate high-level EU AI Act requirements, such as accuracy, robustness, and cybersecurity, into technically measurable, reproducible, and auditable processes tailored to specific intended purposes and operating conditions. Topics include alignment between European harmonized standards and international (ISO/IEC) efforts, jurisdictional differences (e.g., biometrics and fundamental rights), and the role of standards in building customer trust.

In this episode, hosts Michael Mainelli and Adam Leon Smith welcome Cindy Parokkil, ISO’s AI Policy Lead and head of the Standards and Public Policy Programme, for a wide-ranging discussion on the evolving relationship between AI standards, governance, and public policy. Drawing on her experience at the United Nations, BSI, and now ISO, Cindy explains how international standards help bridge the gap between policy ambitions and practical implementation, enabling organisations and governments to turn high-level AI principles into real-world processes, risk management frameworks, and governance tools.

The conversation explores the landmark Seoul Statement launched by ISO, IEC, and ITU, examining its four core commitments: socio-technical approaches to AI, the relationship between standards and human rights, multi-stakeholder collaboration, and capacity building. Cindy discusses why these themes have become increasingly important as governments, regulators, industry, and civil society seek common approaches to trustworthy AI governance.

The episode also looks inside ISO/IEC JTC 1/SC 42, the international committee responsible for AI standards, highlighting efforts to increase participation from developing nations and strengthen links between policymakers and technical experts. Cindy shares examples from countries including Malaysia, South Africa, Brazil, and the UK, illustrating how standards can support national AI strategies and governance objectives. A thoughtful exploration of how international standards bodies are adapting to the challenges of AI, fostering global collaboration, and ensuring that diverse voices help shape the future of AI governance.

In this episode, hosts Michael Mainelli and Adam Leon Smith welcome James Gealy, Standardization Lead at Safer AI, a French NGO focused on AI risk modelling. With a background in spacecraft testing and operations at Northrop Grumman and Airbus, James brings hard-earned lessons from high-stakes risk and quality management, inspired by the Challenger disaster, into the world of AI standards.

James discusses Safer AI’s pioneering risk modelling work on how frontier models like Claude’s Mythos could amplify cyber threats, benchmark saturation, and real-world harms. He also shares insights on his key standards roles, including editing ISO/IEC 42119 Part 8 on LLM benchmarking and red teaming, co-editing the European AI Risk Management Standard, and leading a new Preliminary Work Item on the safety of AI models, systems, and applications.

The conversation explores balancing innovation with safety, standards’ role in the EU AI Act, international collaboration, and the challenges of addressing frontier risks. A timely, pragmatic discussion on turning real-world risk insights into actionable global standards.