Phillip Rieger
Dr.-Ing. B.Sc. M.Sc.Secure Systems, Assistant Professor
Phillip Rieger is an Assistant Professor for Secure AI Systems at ISEC. His research focuses on the security and privacy of AI systems as well as resilient applications of AI to cybersecurity.
Before joining TU Graz, he received his PhD in Computer Science from TU Darmstadt in 2025. His research spans the security and privacy of distributed learning, security and reliability of generative AI systems, and the use of AI for cybersecurity. His work is frequently published at top-tier security venues.
Join my Team
I am looking for PhD students who share my passion for AI security and want to understand why AI systems are vulnerable, how these vulnerabilities can be exploited, and how we can fix them. Given the widespread deployment of AI, this research spans a broad range of topics, including the security and privacy of distributed learning (e.g., Federated Learning and Split Learning), the security and reliability of LLMs (e.g., prompt injection, agent security, and safety alignment), and deepfakes.
We also investigate how AI can be utilized to secure systems and what new challenges arise when AI moves from the ML sandbox into the security domain, for example, when AI is used to detect network and system intrusions, discover vulnerabilities, or investigate privacy implications.
A PhD is your entry point into research. You will be introduced to all aspects of academic research: developing new attacks and defenses for AI systems, turning ideas into research projects, presenting and publishing your work at top international conferences, contributing to research projects, and supervising and teaching students to inspire the next generation of researchers in AI security.
If this matches your research interests, feel free to reach out with your CV, transcript of records, and diplomas, as well as your Master’s thesis if already available.
Join my Team
I am looking for PhD students who share my passion for AI security and want to understand why AI systems are vulnerable, how these vulnerabilities can be exploited, and how we can fix them. Given the widespread deployment of AI, this research spans a broad range of topics, including the security and privacy of distributed learning (e.g., Federated Learning and Split Learning), the security and reliability of LLMs (e.g., prompt injection, agent security, and safety alignment), and deepfakes.
We also investigate how AI can be utilized to secure systems and what new challenges arise when AI moves from the ML sandbox into the security domain, for example, when AI is used to detect network and system intrusions, discover vulnerabilities, or investigate privacy implications.
A PhD is your entry point into research. You will be introduced to all aspects of academic research: developing new attacks and defenses for AI systems, turning ideas into research projects, presenting and publishing your work at top international conferences, contributing to research projects, and supervising and teaching students to inspire the next generation of researchers in AI security.
If this matches your research interests, feel free to reach out with your CV, transcript of records, and diplomas, as well as your Master’s thesis if already available.
Research
Our team focuses on the intersection of security and AI. AI systems have become an integral part of modern life, are increasingly entrusted with sensitive and critical tasks, and are often relied upon without questioning their outputs. We investigate attacks against AI systems, their underlying causes and requirements, and design effective defenses. We study security and privacy across different areas of AI, from federated learning to generative AI, with the goal of making AI systems secure and privacy-preserving by design.At the same time, AI can not only analyze complex data but also analyze complex data at scale, making it a valuable tool for cybersecurity. We develop resilient methods for applying AI to security problems, taking into account the unique challenges of the domain, such as scarce attack data, untrusted inputs, adaptive attackers, and intentional manipulation.
Teaching
AI Security and Privacy (705.224, Winter term)A research-oriented seminar in which students explore the literature on a current problem in AI security and privacy, systematically categorize and compare existing work, and present their findings in a scientific report and presentation.
Security and Privacy of Distributed Learning (705.223)
This course focuses on security and privacy of distributed and federated learning, including attacks, privacy threats, and mechanisms for secure and privacy-preserving collaborative learning.
If you have a passion for AI security and want to go beyond simply using AI, there are always exciting research problems to explore. AI systems are being deployed everywhere, but we are only beginning to understand their security. What happens inside these systems? Where are their limits? How can they be broken or manipulated, and how can we build robust defenses? I regularly offer Bachelor’s and Master’s thesis topics exploring these and related questions.
Service
Peer-reviewing as a Program Committee Member:* Network and Distributed Systems (NDSS) Symposium 2027
* ACM ASIA CCS 2027