CISPA Helmholtz-Zentrum für Informationssicherheit

Whether in the cinema or through exchange with the next generation of scientists – July showed the many different ways research at CISPA can be experienced. 🍿 At the CISPA Cyber Cinema, the cult classic Blade Runner was in the spotlight. Together with Prof. Dr. Jilles Vreeken, CISPA Faculty and expert in machine learning, the audience discussed which technologies portrayed in the movie have already become reality and what opportunities and challenges artificial intelligence brings for our society. 🎓 Our next generation of researchers also provided insights into the research of tomorrow: During the Bachelor's & Master's Thesis Intro Poster Session, students presented their planned theses and shared first ideas for their research projects. The poster session created space for exchange, valuable feedback, and new impulses, while highlighting the diversity of research at CISPA. Thank you to everyone involved and all visitors who made these events possible. We are looking forward to many more opportunities to make research accessible and experience it together.

Some researchers ask questions that deepen our understanding of the digital world. Sometimes, those same questions lead to another one: Could this research also become a technology that people rely on every day? Research creates knowledge. It advances science, shapes future discoveries, and often lays the foundation for innovations we cannot yet predict. In some cases, it also opens the door to technologies that address real-world challenges directly. At CISPA, we see technology transfer as one way—alongside scientific excellence—to maximize the impact of research. It creates opportunities for researchers who want to translate their findings into products and services that strengthen cybersecurity and trustworthy AI. One example is the Founders Fellowship from our CISPA Technology Transfer, which supports researchers exploring this entrepreneurial path. Our Fellow, Kevin Morio, is building on his research in formal methods to develop technology that continuously verifies whether security-critical software behaves exactly as intended. While security protocols can be mathematically proven secure, their implementations may still behave differently in practice. His startup idea aims to detect these deviations in real time, helping bridge the gap between verified designs and deployed software. Whether through publications, open science, collaborations, or startups, the goal remains the same: ensuring that excellent research creates lasting impact. Interested in our full interview with Kevin? Read it here: https://lnkd.in/gt488jc3

Award Honors Research with Lasting Impact 🎉 CISPA-Faculty Prof. Dr. Cas Cremers has received the IEEE Computer Security Foundations Symposium (CSF) 2026 Test-of-Time Award together with Prof. David Basin (ETH Zurich), Dr. Benedikt Schmidt, and Dr. Simon Meier! 🏆 Their 2012 paper, "Automated Analysis of Diffie-Hellman Protocols and Advanced Security Properties," has been recognized for its enduring significance and outstanding impact. Since its publication, the methods introduced in the paper have influenced numerous subsequent research efforts in formal security analysis. The researchers implemented their analysis approach in the Tamarin Prover, whose core algorithm is described in the award-winning paper. Over the years, the tool has been continuously extended and is now regarded as one of the world's leading tools for modeling and analyzing cryptographic protocols. The paper was written while all four researchers were members of David Basin's Information Security Group at ETH Zurich. Today, Cas Cremers is a CISPA-Faculty, Benedikt Schmidt works at Snowflake und Simon Meier is with Digital Asset. Congratulations to all award recipients! Learn more at: https://lnkd.in/dy3XCKRj

The incident involving OpenAI and Hugging Face has sparked widespread discussion. For us, the individual case is less important than the fundamental research questions it brings to light, as CISPA-Faculty Lea Schönherr depicts it. Powerful AI systems open up new possibilities. They can identify vulnerabilities faster and, in doing so, help us make systems more secure. At the same time, they can pursue goals in ways we do not expect. That is why it is not enough to ask whether a model can solve a task. We also need to understand how it goes about solving it and the conditions under which it operates. This is exactly where our research on trustworthy AI comes in: How can we develop and evaluate AI systems so that they are powerful, understandable, and safe to use?

What a spectacular finale! After six regional qualifiers across Europe, the first CISPA European Cybersecurity & AI Hackathon Championship reached its thrilling conclusion in early July, where the champions were crowned. Congratulations to all finalists for their outstanding performances, and especially to the winning teams on this remarkable achievement. Your talent, dedication, and innovative spirit made this championship an unforgettable experience. 🎥 Watch some of the highlights from the championship final in the video below.

CISPA Researcher Warns Against One-Size-Fits-All Age Verification Online Age assurance technologies (AATs) have recently attracted significant public attention in debates about online child protection. But what do we know about their effectiveness, potential side effects, and public acceptance? We spoke with CISPA-Faculty Dr. Wouter Lueks about this topic and asked him about the trade-offs these technologies present for privacy, security, accessibility, and civil liberties. Together with colleagues from the University of Hamburg and the Leibniz Institute for Media Research, Lueks conducted a comprehensive analysis of age assurance technologies and provides recommendations on which types of AATs are more—or less—suited to protecting minors online. Their findings highlight that age assurance systems can have significant implications for privacy, accessibility, and civil liberties, and that no single approach is suitable for all contexts. How can we balance child protection with privacy and civil liberties online? Read what Dr. Wouter Lueks suggests: https://lnkd.in/egSaddMX Leibniz-Institut für Medienforschung | Hans-Bredow-Institut

Excellent research creates new knowledge. Technology transfer helps explore how that knowledge can create impact beyond academia. At CISPA, both are part of our mission. CISPA Founders Fellow Alfusainey Jallow spent years researching how code fragments are reused and evolve across modern software ecosystems. His work revealed a challenge with real-world implications: organizations often lack visibility into how related code fragments develop over time—and how important information such as security updates, bug fixes, or licensing changes may fail to reach connected fragments. Research made that insight possible. Now, through the CISPA Founders Fellowship, Alfusainey is exploring how this research can become a practical solution. With FragmentIQ, he is investigating new approaches to help organizations better understand the relationships between code fragments across today's software ecosystem. This is what technology transfer means: building on scientific excellence and exploring new ways for research to create impact in society and industry. Follow CISPA Technology Transfer to see more research come to life!

How to Tackle Child Sexual Abuse Material Generation by Text-to-Image Models Child sexual abuse material (CSAM) has long been a serious and persistent problem. However, the emergence of AI image generators has introduced an alarming new dimension: the ability to generate synthetic CSAM. One commonly proposed safeguard is to remove images of children from the training datasets used to develop these models. CISPA-Faculty Dr. Ana-Maria Cretu investigated how effective this approach is and what its consequences are. Her findings show that filtering child images from training datasets provides only limited protection and can be circumvented. Filtering can also introduce unintended biases in model behavior. Overall, Cretu's findings suggest that dataset filtering alone is not sufficient and that preventing AI-generated CSAM requires more comprehensive safety measures. Learn more in the full article below.

Alumni Spotlight: We caught up with Sahar Abdelnabi, who was just appointed to the EU AI Office’s Scientific Panel, alongside former supervisor Professor Dr. Mario Fritz. If you completed a degree with CISPA you can sign up to our newsletter: https://lnkd.in/dU-Vfs4Q