Bezpečnost AI a její zodpovědné využítí ve firemním prostředí

Knowledge within the scope of the courses listed in the Previous courses and Related courses
sections

• Introduction to AI - how it works and what to expect from it
AI as a statistical model - why it is not an "omniscient system"
Input quality = output quality (principles of correct prompting)
Critical reading of outputs - the need to verify facts and security content

• Risks and regulatory frameworks
Data types: public, internal, protected, strictly protected – what to never put into AI
GDPR and personal data protection in relation to AI
Other legislation (AI ACT, SK law – in preparation)
Real incidents (leaks of business plans, source code, personal data)
Impacts on the legal status and reputation of the company
Moral requirements and the consequences of their non-compliance

• Comparison of AI tools from a security perspective
Geopolitical view of AI tools (USA, EU, China, …)
ChatGPT, Perplexity, Google Gemini, Microsoft Copilot – differences in data protection
Enterprise AI and their data work mode (Microsoft, Google, OpenAI Enterprise)
On-premise / offline AI solutions – when and why to deploy them as a “plan B”

• Safe use of AI in practice
AI as a source of data leakage
Where does it end data
What is in the prompt and output, anyone can see
Data minimization and anonymization when working with AI
How to properly set up the environment and leave a minimum of digital traces
What the tool remembers and how to manage data retention and history
Knowledge and Library Creation (GPT)
Using only approved AI tools and checking settings (data protection mode)
Prompt engineering. How to form prompts, what basic parameters, specifications, structure prompts should have
Practical exercises: secure prompts, extraction and analysis of documents without the risk of data leakage

• Effective use and procedures
Best practices rules and recommendations for employees
Creating your own models, agents
Chat management
The relationship between chat, project and agent
Using Knowledge and Libraries
Effective use - keyboard shortcuts, examples of contraindications
Knowledge on how to effectively and safely use AI tools in everyday work
Managing the structure of chats
Teamwork and corporate sharing of AI use