Open Topics

AI systems increasingly interact with tools and sensitive data. This topic investigates how access control, permission models, and policy enforcement mechanisms can be integrated into such systems.

Systematically analyse existing evasion attacks and defenses and develop new attacks/defenses in specific application domains (such as industrial productions systems).

• Building an LLM agent that drives Ghidra/IDA/radare2 to answer reverse-engineering questions autonomously
• Task decomposition and tool-use strategies for multi-step binary analysis (unpacking, deobfuscation, semantic recovery)
• Benchmarking agent performance against human reverse engineers on curated…

• Authentication, authorization, and accountability for autonomous agents acting across systems
• Secure delegation and capability-based access control for AI agents
• AI agent consensus

Please contact Aljosha Judmayer.

Organizations must comply with security standards and regulations. This topic explores how AI-based systems can support compliance processes and security governance.

• Analysing the Android implementation
• Bypassing the permission in iOS through webviews

Please contact Sebastian Schrittwieser.

Detecting anomalies in complex systems is a key cybersecurity challenge. This topic investigates how AI-based systems can support anomaly detection.

Analyse and evaluate attack vectors for poisoning attacks, evaluate their effectiveness and side-effects, as well as existing defenses, and develop new detection/defense mechanisms.

• Designing LLVM passes that inject self-checking guards at compile time
• Automating placement of tamperproofing checks based on static analysis of sensitive code regions

Please contact Sebastian Schrittwieser.