EU Cyber Acts Conference 2025: A Deep Dive into Securing AI

By Omar Santos, Distinguished Engineer, Cisco

On March 25, 2025, the EU Cyber Acts Conference in Brussels will bring together cybersecurity professionals, policy-makers, and industry leaders to discuss one of the most pressing challenges of our time—securing artificial intelligence. Several members of the Coalition for Secure AI (CoSAI) will be participating and presenting at the conference. A key highlight of the conference is the “AI Cyber Day” track, which zeroes in on the global evolution of cybersecurity certification frameworks tailored for AI systems.

Securing AI Applications and Agentic Systems

I’m truly honored to be among such a distinguished lineup of speakers, presenting “Securing AI: Navigating Security Challenges in Modern AI Implementations.” I am excited to share insights on the evolving threats and best practices for securing AI systems! In this session, we will unpack the layered security considerations inherent to today’s AI implementations. Attendees will be guided through the best practices surrounding AI operations (AI Ops), model development, fine-tuning, and deployment of AI applications. Emphasis will be placed on well-known techniques such as Retrieval Augmented Generation (RAG) and how to secure innovative agentic systems (methods that represent the forefront of modern AI security strategies).

The session also introduces the Coalition for Secure AI (CoSAI), an open project that is bringing experts from industry-leading organizations dedicated to sharing best practices for secure AI.  

CoSAI’s focus is to better equip the community to fortify the AI supply chain, equip defenders for emerging threats, secure agentic AI systems, and promote robust security risk governance frameworks. As the digital ecosystem evolves, CoSAI will help organizations of all sizes to secure their AI implementations.

Both the EU and the US are spearheading efforts to create harmonized regulatory frameworks that set common frameworks, methodologies, and promote innovation. There is a common realization that we all need to have a balance that doesn’t hinder innovation. During AI Cyber Day, experts will discuss:

• The current state of regulatory development in both regions.
• Future outlooks for AI security regulation.
• Best practices for robust risk management and promoting innovation.

Collaborative Panel Discussion: Charting the Future of AI Security

Complementing the technical presentations, the conference will feature a panel discussion titled “Collaborative Efforts to Secure AI and AI Applications and Services.” In this session, my colleague Piotr Ciepiela, from EY, will highlight the role of CoSAI in advancing secure AI practices.

This session will bring together influential voices from governments, industry, and academia to share insights on building a safer AI ecosystem. The discussion will focus on:

• Developing and sharing best practices and tools for secure AI deployment.
• Collaborative strategies to navigate the challenges of securing AI applications.
• The role of CoSAI in uniting diverse stakeholders under a common mission of enhanced AI security.
  

The panel includes a distinguished lineup of speakers:

• (Moderator) Matthias Intemann, Head of Digitisation, Bundesamt für Sicherheit in der Informationstechnik (BSI), Germany
• Franziska Weindauer, CEO of TÜV AI.Lab, TÜV-Verband, Germany
• Piotr Ciepiela, Partner and EMEIA Cybersecurity Leader at EY, Poland
• Ezi Ozoani, Representative to GPAI Code of Practice at Applied AI Institute for Europe GmbH, Ireland           .

Looking Forward to the Discussion

The spotlight on AI security is more intense than ever. With evolving threats and rapid technological advancements, the EU Cyber Acts Conference offers invaluable insights for anyone involved in the AI space.

Whether you’re an industry veteran, a policymaker, or simply curious about the future of AI, these presentations are set to provide great insights to what’s new in AI and cybersecurity. Mark your calendars for March 25-26, 2025, and join the conversation on securing the future of AI.

Additional Exciting News

CoSAI is part of the OASIS Open ecosystem where many other related technical initiatives are happening. A new one worth noting is the Data Provenance Standards (DPS) Technical Committee which links to a shared focus on ensuring trust and security of AI systems, as well as great business outcomes. DPS will develop a standardized metadata framework for tracking data origins, transformations, and compliance, helping organizations establish clearer governance practices. Visit the OASIS Open website to learn more about this new initiative and others.