AI - Your New Advantage or Your Next Risk?

What has changed over the past year? The first edition focused on the question of how to deploy AI safely. Today, organizations are facing the next stage: AI is gaining access to data, acquiring permissions, and beginning to take action.

A brief introduction to the second edition, outlining the questions companies will need to address in 2027.

Practical Use of Microsoft AI Agents in Threat Detection and Incident Response.

Artificial intelligence has changed the rules of the game for both sides. Attackers are using it to act faster, more cheaply, and on a larger scale, and a traditional security model based solely on human effort can no longer keep up.

In this live session, Microsoft experts will demonstrate how AI agents support organizations in their day-to-day security efforts: from threat detection, through incident analysis, to responding to an attack before it causes harm. You’ll see the latest Microsoft Security solutions, including Security Copilot, the Perception platform, and Agent 365 - presented in Poland for the first time.

You'll learn what human-AI collaboration looks like in practice, which of these capabilities are already included in your existing Microsoft 365 licenses, and where to start so your organization can defend itself at the same speed it's being attacked.

Employees are using AI—with or without your approval. Bans don't work: they simply lead to "Shadow AI," the uncontrolled transmission of company data to external models. The result? A risk of confidential data leaks and a loss of control over information.

In this session, we'll show how to turn bans into control. We'll cover the key risks of uncontrolled AI use: Shadow AI, loss of data control, and excessive trust in models.

We'll also present our proprietary solution for using LLMs safely: employees can continue to use any model available online, while confidential and sensitive data never leaves the organization.

Instead of blocking AI, make it available safely, within an auditable, regulation-compliant architecture.

AI has reduced the time it takes to build solutions, but it is also increasing the number of systems, dependencies, and decisions that teams do not always fully understand.

The growing gap between what teams build and what they truly understand, agent memory, and today's architectural decisions could end up being a bill that organizations won't pay until 2027.

Today, the biggest challenge is no longer building the first AI solution. What matters now is securing the right skills, controlling costs, establishing an operating model, and ensuring business continuity so that AI can continue to deliver business value.

In this session, we'll examine the challenges organizations face when moving from a pilot to production, and how to plan for the development, maintenance, and oversight of AI. We'll discuss skills, SLAs, governance, cost, and how to build an operating model that makes AI a permanent part of the organization rather than a one-time project.

Artificial intelligence is increasingly suggesting what we should read, watch, buy, and decide. But it can just as easily mislead us, present false information, and shape the way we think.

The increasing convenience of AI can lead to intellectual laziness and the gradual relinquishment of responsibility for our own choices.

This talk will explore a reality in which AI no longer merely answers questions, but begins to independently make recommendations, decisions, and, often, take action on a person's behalf.

Professor Andrzej Dragan - a physicist at the University of Warsaw and the National University of Singapore, as well as an acclaimed photographer and filmmaker.

As of August 2, 2026, under the EU's AI Act (Regulation (EU) 2024/1689), all entities that use generative AI content are required to label it in the digital space.

Will this labeling requirement give content created by experts an advantage over algorithmic output, and how can these rules be implemented in practice?

In this talk, I'll present specific use cases for the new regulation and demonstrate how to develop effective AI governance policies within companies.

A summary of the key takeaways.

A YouTube stream distributed to registered participants of the ISCG AI 2026 conference on October 8, 2026.