Trustworthy AI series: Privacy and security

An episode of the Changing Conversations Podcast podcast, hosted by SGS, titled "Trustworthy AI series: Privacy and security" was published on March 18, 2025 and runs 47 minutes.

March 18, 2025 ·47m · Changing Conversations Podcast

0:00 / 0:00

Summary

As artificial intelligence continues to revolutionize industries, a critical question remains: how do we keep AI systems secure and protect sensitive data from breaches? In this third episode of our Trustworthy AI series, we explore the unique security and privacy risks that AI presents—and the cutting-edge solutions shaping its future.Join us as we dive into:✅ The real-world consequences of AI security failures and privacy violations.✅ How cryptography and privacy-enhancing technologies are safeguarding data like never before.✅ The evolving regulatory landscape, including the EU AI Act, and what it means for businesses.✅ Strategies to anticipate and mitigate AI-specific threats before they happen.🎙️ Featuring AI and cybersecurity experts:🔹 Willy Fabritius – Global Head of Strategy & Business Development Information Security, SGS🔹 Tomislav Nad – Lead Innovation Technologist, SGS🔹 Andreas Trügler – Co-Head of Data Privacy for AI, Know-CenterWith AI security breaches on the rise and regulations tightening, staying ahead is no longer an option—it’s a necessity. Tune in to gain expert insights and learn how to build trustworthy AI systems that stand the test of time.(00:00:16)- Introduction.(00:06:14)- What privacy and security means in the context of AI(00:09:27)- What possibilities are available to protect sensitive information in the age of AI?(00:16:00)- Is it possible to protect an AI model against adversarial attacks? (00:22:56)- Real-world example(00:28:56)- How trustworthy has to be an organization(00:30:20)- What other attacks exist on AI system?(00:34:47)- Is there a way to evaluate how robust an AI system is in general against malicious attacks? (00:37:00)- What impact could a successful attack have on my business? How do I protect against these kinds of attacks? (00:42:22)-Where do you see the biggest challenges in the field? About our “Trustworthy AI: current areas of research and challenges” series:The need for trustworthy Artificial Intelligence systems is recognized by many organizations, from governments, to industries and academia. As AI systems become more widely used by both organizations and individuals, it is important to establish trust in them. To establish this trust, numerous white papers, proposals and standards have been published and are still in development to educate organizations on the need for and uses of AI systems. Join us for our series as our experts discuss a variety of topics related to building trust and understanding of AI systems.

Episode Description

As artificial intelligence continues to revolutionize industries, a critical question remains: how do we keep AI systems secure and protect sensitive data from breaches?

In this third episode of our Trustworthy AI series, we explore the unique security and privacy risks that AI presents—and the cutting-edge solutions shaping its future.

Join us as we dive into:✅ The real-world consequences of AI security failures and privacy violations.
✅ How cryptography and privacy-enhancing technologies are safeguarding data like never before.
✅ The evolving regulatory landscape, including the EU AI Act, and what it means for businesses.
✅ Strategies to anticipate and mitigate AI-specific threats before they happen.

🎙️ Featuring AI and cybersecurity experts:
🔹 Willy Fabritius – Global Head of Strategy & Business Development Information Security, SGS
🔹 Tomislav Nad – Lead Innovation Technologist, SGS
🔹 Andreas Trügler – Co-Head of Data Privacy for AI, Know-Center

With AI security breaches on the rise and regulations tightening, staying ahead is no longer an option—it’s a necessity. Tune in to gain expert insights and learn how to build trustworthy AI systems that stand the test of time.

(00:00:16)- Introduction.

(00:06:14)- What privacy and security means in the context of AI

(00:09:27)- What possibilities are available to protect sensitive information in the age of AI?

(00:16:00)- Is it possible to protect an AI model against adversarial attacks?

(00:22:56)- Real-world example

(00:28:56)- How trustworthy has to be an organization

(00:30:20)- What other attacks exist on AI system?

(00:34:47)- Is there a way to evaluate how robust an AI system is in general against malicious attacks?

(00:37:00)- What impact could a successful attack have on my business? How do I protect against these kinds of attacks?

(00:42:22)-Where do you see the biggest challenges in the field?

About our “Trustworthy AI: current areas of research and challenges” series:
The need for trustworthy Artificial Intelligence systems is recognized by many organizations, from governments, to industries and academia. As AI systems become more widely used by both organizations and individuals, it is important to establish trust in them. To establish this trust, numerous white papers, proposals and standards have been published and are still in development to educate organizations on the need for and uses of AI systems. Join us for our series as our experts discuss a variety of topics related to building trust and understanding of AI systems.