{"id":354,"date":"2026-07-27T12:33:25","date_gmt":"2026-07-27T10:33:25","guid":{"rendered":"https:\/\/it-defense.de\/it-defense-2027\/?page_id=354"},"modified":"2026-07-27T12:33:26","modified_gmt":"2026-07-27T10:33:26","slug":"artificial-intelligence-and-it-security","status":"publish","type":"page","link":"https:\/\/it-defense.de\/it-defense-2027\/en\/artificial-intelligence-and-it-security\/","title":{"rendered":"Artificial Intelligence and IT Security"},"content":{"rendered":"\n

Artificial Intelligence and IT Security \u2013 from Risk Assessment to AI Policy<\/h1>\n\n\n\n

Instructor:<\/strong> Felix Friedberger<\/a><\/p>\n\n\n\n

Duration:<\/strong> 1 Day \u2013 January 26, 2027<\/p>\n\n\n\n

Content:<\/strong>
This one-day seminar covers the IT security aspects of using AI systems in enterprises \u2013 with the aim to equip you with the knowledge necessary to make informed IT governance decisions and create your own AI policy.<\/p>\n\n\n\n

AI and AI-based tools have quickly found their way into nearly every area of the company: developers use coding assistants, specialist departments use chatbots and Office integrations, sensitive data is made searchable using AI, and more and more often AI features are built into existing products or AI-based applications are bought. Understanding these use cases, assessing their risks realistically and transferring them into robust regulations poses a challenge for security executives. Without a basic understanding of the underlying technology \u2013 from LLM hosting providers to tool usage and the typical attack surface \u2013 neither the requests of individual employees nor the purchase of entire products can be evaluated appropriately.<\/p>\n\n\n\n

At the beginning of the seminar, we will teach you the technical fundamentals required for reliable risk assessment: tokenization and inference, the path of data from the tool through the API to the model, the concept of MCP, agents and tool usage, and the OWASP LLM Top 10 as an evaluation framework. Simple examples are used to show where data can leak, how tool calls can be malicious and what impact traditional attacks like prompt injection can have.<\/p>\n\n\n\n

Based on this, we will look at the different use cases and their respective risks: the private and often unnoticed use of public AI services (\u201cshadow AI\u201d), using AI applications on workstations and their extensive user context permissions, the special risks with developer tools like coding agents, and creating in-house AI applications with RAG, identity and permission questions. We will also categorize publicly accessible AI applications and the relevant legal conditions, in particular the EU AI Act and data protection issues. The scenarios are explained using concrete examples and demonstrations.<\/p>\n\n\n\n

Finally, we will combine the things we have learned and translate them into practical governance by jointly working on typical example cases from everyday business situations \u2013 but now with the knowledge gained in the seminar.<\/p>\n\n\n\n

After the seminar, participants will not only know how AI systems technically work in general as well as the attack surface and data protection risks they entail, but also how to evaluate their employees\u2019 requests for AI usage and develop an acceptable AI policy for their company.<\/p>\n\n\n\n

Topic areas:<\/strong><\/p>\n\n\n\n