Information Technology and Artificial Intelligence Security
Niveau
Second cycle of study, master's degree
Learning outcomes of the courses/module
The participants:
• can explain the fundamentals of IT security, including threat models and the CIA triad
• can describe typical cybersecurity threats and explain basic protective mechanisms
• can describe the specific characteristics of AI security and analyze AI-specific attack vectors (prompt injection, data poisoning, model inversion, adversarial attacks)
• can assess the risks and failure modes of generative models (LLMs) and derive appropriate security, monitoring, and governance measures
• can securely apply AI systems within IT infrastructures
• can explain the fundamentals of IT security, including threat models and the CIA triad
• can describe typical cybersecurity threats and explain basic protective mechanisms
• can describe the specific characteristics of AI security and analyze AI-specific attack vectors (prompt injection, data poisoning, model inversion, adversarial attacks)
• can assess the risks and failure modes of generative models (LLMs) and derive appropriate security, monitoring, and governance measures
• can securely apply AI systems within IT infrastructures
Prerequisites for the course
Foundational knowledge of programming and mathematics/statistics
Course content
IT Security Fundamentals
- Security objectives (CIA), threat models, risk assessment
- Authentication, authorization, identity management
- Fundamentals of cryptography (hashing, symmetric/asymmetric methods)
Security in Machine Learning and AI Systems
- Characteristics of modern AI systems
- AI-specific attacks:
- Prompt injection, indirect prompt injection
- Jailbreaking and hallucinations
- Data poisoning and backdooring
- Adversarial examples
- Model extraction and model inversion
- Attack surfaces in LLMs, agentic AI, and RAG systems
- Protective mechanisms: guardrails, red teaming, input validation, sandboxing
Secure GenAI and LLM Integration
- API security and access control
- Data and metadata security
- Logging, monitoring, and audit trails
- Content filtering, safety layers, and moderation
- Security objectives (CIA), threat models, risk assessment
- Authentication, authorization, identity management
- Fundamentals of cryptography (hashing, symmetric/asymmetric methods)
Security in Machine Learning and AI Systems
- Characteristics of modern AI systems
- AI-specific attacks:
- Prompt injection, indirect prompt injection
- Jailbreaking and hallucinations
- Data poisoning and backdooring
- Adversarial examples
- Model extraction and model inversion
- Attack surfaces in LLMs, agentic AI, and RAG systems
- Protective mechanisms: guardrails, red teaming, input validation, sandboxing
Secure GenAI and LLM Integration
- API security and access control
- Data and metadata security
- Logging, monitoring, and audit trails
- Content filtering, safety layers, and moderation
Recommended specialist literature
- Anderson, R. (2021): Security Engineering
- Papernot, N. (2018): Security and Privacy in Machine Learning
- Chio, C., Freeman, D. (2018): Machine Learning and Security: Protecting Systems with Data and Algorithms
- Zenker, P. (2026): Genai Security: Secure Chatbots and Agent Systems
- Papernot, N. (2018): Security and Privacy in Machine Learning
- Chio, C., Freeman, D. (2018): Machine Learning and Security: Protecting Systems with Data and Algorithms
- Zenker, P. (2026): Genai Security: Secure Chatbots and Agent Systems
Assessment methods and criteria
Exam
Language
German
Number of ECTS credits awarded
4
Semester hours per week
Planned teaching and learning method
- Lecture with discussion
- Completion of practice exercises
- Interactive workshop
- Completion of practice exercises
- Interactive workshop
Semester/trimester in which the course/module is offered
3