Joint project with Fraunhofer AISEC (Cognitive Security Technologies), TU Berlin (Prof. Dr. Konrad Rieck), CISPA (Prof. Dr. Mario Fritz, Thorsten Holz) , Aleph Alpha, until 2027

Project page: https://www.forschung-it-sicherheit-kommunikationssysteme.de/projekte/aigency

Methods of generative AI now permeate almost all areas of life. Especially since the release of ChatGPT by OpenAI in November 2022, the capabilities of current AI systems have come into the focus of the general public. The application areas of existing models as the basis for countless applications within cybersecurity are far from fully identified from today's perspective but already hint at their high and increasing importance for cybersecurity. For instance, generative AI can help detect attacks, identify and close security gaps in networks. Conversely, these very capabilities can also be used by attackers to penetrate networks more efficiently or to contextually generate (meta)passwords and malware. In terms of spreading and detecting deceptive content like phishing, AI opens up entirely new possibilities, both positive and negative. Moreover, the methods of generative AI are continuously being developed and applied to more areas within cybersecurity. The resulting novel threat scenarios are neither known nor quantified. Yet, assessing such future risks is urgently necessary to target research and development toward the biggest future challenges in AI and IT security. At the same time, the insights gained also contribute to a better fundamental understanding of generative AI, e.g., towards improved causal inference capabilities, which are still very limited in current systems.

Publications:

tba