Artificial intelligence (AI) promises to revolutionize healthcare, but it poses risks from cyberattacks that can compromise patient safety, medical devices, and the work of emergency responders. The “SecureNeuroAI” project, a collaboration between the University of Bonn, the University Hospital Bonn and the FIZ Karlsruhe, is developing secure, AI-supported methods for real-time detection of medical emergencies, using epileptic seizures as an example. The project is funded by the Federal Ministry of Research, Transfer and Space (BMFTR) with around 2.5 million euros over three years. The results are to be applied beyond epilepsy seizure detection and strengthen the cybersecurity of medical systems.

The project, led by Prof. Dr. Elena Demidova, head of the “Data Science and Intelligent Systems” working group at the University of Bonn, focuses on the development of data authentication methods that do not interfere with AI data processing, but prove its origin. Prof. Dr. Michael Meier, head of the “IT Security” working group at the University of Bonn, Prof. Dr. Björn Krüger from the Department of Epileptology at the University Hospital Bonn and Prof. Dr. Franziska Boehm from FIZ Karlsruhe are also involved. The researchers use multimodal sensor data, such as wearables that record vital parameters such as heart and respiratory rate, as well as clinical patient data, to detect seizures in real time.

A central goal is the development of cyber-secure AI models that reliably detect manipulation of sensor data. In parallel, technical, organizational and legal measures are being developed to enable the integration of these AI methods into clinical and home applications. Prof. Demidova’s research group focuses on explainable AI models and tamper detection, while the University Hospital Bonn is responsible for clinical validation and data collection. Prof. Meier contributes expertise in IT security, and FIZ Karlsruhe analyzes data protection and legal issues in order to derive recommendations for digital infrastructures.

The results of “SecureNeuroAI” are intended to protect AI models and data from manipulation and increase the integrity and reliability of AI-based medical devices. In doing so, the project lays the foundation for safer, AI-supported health technologies that go far beyond epilepsy seizure detection.

Editorial office: X-Press Journalistenbüro GbR

Gender note. The personal designations used in this text always refer equally to female, male and diverse persons. Double/triple references and gendered designations are avoided for the sake of better readability ected.