Episode 12: The Role of the TIC Industry in Assessing AI

In this episode of the ‘TIC for AI’ series, we explore the growing importance of AI regulation and its impact on the testing, inspection and certification (TIC) industry. Hosted by Tomislav Nad, Lead Innovation Technologist at SGS, the episode features two prominent guests: Jeremy Rollison, Head of EU Policy at Microsoft, and Gabriele Mazzini, Architect and Lead Author of the AI Act. They discuss the evolving role of the TIC industry in AI governance, the complexities of certifying cutting-edge AI technologies and how international standards are shaping compliance efforts across the globe. Join us for insights into the TIC sector’s critical role in ensuring responsible AI development and deployment.

Listen to the podcast

About the speakers

Join our experts as they delve into the crucial role of the TIC industry in AI regulation and certification:

• Tomislav Nad, Lead Innovation Technologist, SGS
• Jeremy Rollison, Head of EU Policy, Microsoft
• Gabriele Mazzini, Architect and Lead Author of the EU AI Act

Podcast highlights

What is the role of the TIC industry in AI governance?

The TIC industry plays a vital role in AI governance by ensuring that systems are assessed for compliance before entering the market. The EU AI Act expands TIC’s traditional responsibilities, such as health and safety, into areas like bias and discrimination. This presents challenges, particularly in upskilling talent, but is crucial for responsible AI development. The Act balances third-party assessments with self-assessments, ensuring thorough compliance while allowing the industry time to adapt.

How does the EU AI Act address the rapid development of AI technologies?

The AI Act highlights the importance of international standards to manage AI risks, extending beyond just AI models to organizational structures. Drawing on principles like "Know Your Customer" from financial services, the Act encourages responsible AI development. Third-party certifications ensure that companies comply with standards, which helps streamline compliance across jurisdictions. This global alignment supports ethical AI practices and makes it easier for organizations to meet regulatory demands.

What are the challenges of certifying AI systems with multiple components from different suppliers?

Certifying AI systems that involve components from various suppliers is complex. The AI Act holds the end provider responsible for ensuring that all integrated components meet regulatory standards. This is managed through contractual agreements between providers and suppliers. While the current framework relies on these contracts, there is potential for future component certification schemes. Ultimately, the goal is to ensure that high-risk AI applications comply with regulations without overwhelming providers.

About the Changing Conversations podcast

Join us for the Changing Conversations podcast, where we take a deep dive into the complex issues facing the modern world and shine a light on the innovations that can make a difference. Each episode offers a fresh perspective on a unique topic – from the transition to a more sustainable world to AI technology.