The integration of AI into safety-critical applications poses new requirements for cybersecurity and real-time capability. A particularly challenging task is the development of fault-tolerant and trustworthy AI systems for safety-critical environments.
The KIbSIS project (AI-based Security for Industrial Systems) addresses these challenges through innovative cybersecurity methods. The core of the project is the development of an AI-based solution for automatic acknowledgment of safety devices on machine tools in smart factories while complying with safety requirements as well as regulations for AI and cybersecurity. AI-based anomaly detection in networks and industrial systems identifies disruptions and cyberattacks at an early stage and enables targeted countermeasures. Additionally, the project develops crypto-agile security mechanisms with post-quantum-secure algorithms on secure elements as well as migration strategies for cryptographic solutions.
