Anna Goławska

These days, it takes just a few hours to come up with an idea for using AI. It’s much harder to build a solution that’s secure, scalable, accepted by users, and actually put to use within an organization.

We will devote this panel to the practical aspects of AI implementation. We will discuss the experiences of companies that have progressed from experiments and proof-of-concepts to production-ready solutions, as well as
the reasons why many projects never move beyond the pilot phase. It’s not just the successes that matter, but also the mistakes, failed experiments, and lessons learned from
real-world implementations.

Topics of discussion include:
 Why does a PoC work, but production deployment doesn’t?
 What mistakes do companies make most often, and what would they not do again today?
 How to identify areas with the greatest potential for AI adoption,
 How to move from individual experiments to AI operating at an organizational scale,
 In-house models, open-source solutions, or off-the-shelf platforms?
 how to prepare data, processes, and infrastructure,
 how to assess the quality and reliability of AI-based solutions,  who should own AI implementations: business, IT, or a dedicated AI team?
 what really determines whether an AI project moves beyond the pilot phase and begins to create value across the entire organization

AI is changing the rules of the game in cybersecurity. It enables faster threat analysis, anomaly detection, and automated incident response. At the same time, it gives attackers the tools to create more convincing scams, automate attacks, and operate on a scale never seen before.

On top of that, there’s a new challenge: we must not only use AI to protect
organizations, but also protect the AI itself—the models, agents, data, and systems to which
they have access.

During the panel, we will examine this change from the perspectives of both attackers and defenders.
Topics for discussion include:

• Who makes greater use of AI today—cybercriminals or security teams,
• AI-powered phishing, deepfakes, and the next generation of social engineering,
• new areas of risk,
• new attack types,
• Shadow AI and the risk of uncontrolled use of company data,
• how to secure models, data, and applications that use LLMs,
• identity and trust in a world where sound, images, and video can be generated,
• Where should the line be drawn regarding AI autonomy in the area of security,
• How will the role of cybersecurity specialists change?
• Will AI give a greater advantage to defenders or attackers—and what will cybersecurity look like as autonomous systems carry out more and more actions on both sides?