Selected Topics of Information Security (AI Security and Privacy) (WS 2026)

Content

Artificial Intelligence is increasingly deployed in real-world applications, including systems handling critical and sensitive tasks. However, the widespread adoption of AI, especially for critical tasks, also raises important concerns regarding its security, privacy, robustness, and reliability. This seminar introduces students to current research topics in AI security and privacy. Students work on assigned topics based on recent scientific literature and explore different aspects of securing AI systems and understanding their limitations. Example topics include security and privacy attacks and defenses, robustness in distributed learning, backdoor and data poisoning attacks, fairness in machine learning, inference-time attacks, security challenges of Large Language Model (LLM) agents, hallucinations and reliability issues in LLMs, adversarial attacks against machine learning systems, privacy-preserving machine learning, and the application of AI techniques for security tasks.

Administrative Information

Topics will be assigned after the kick-off meeting. Students will work in groups of up to three. Each group is expected to submit an intermediate report and a final report, and to present their work at the end of the seminar. All course-related communication and announcements will take place via TeachCenter

Lecture Dates

Date	Begin	End	Location	Event	Type	Comment
2026/10/16	11:00	12:00	CCGEG002	CLASS	SE	CLASS/

Lecturers