Safety-Critical Data Infrastructures

In the research field of safety-critical data, we work on the development of secure data infrastructures. The focus is on method and technology bundles which offer both protection and the authorised use of sensitive data in distributed data infrastructures. How can I enable the availability of data for all legitimate users and at the same time guarantee the data sovereignty? How do I determine data quality and eliminate e.g. cognitive biases in data sets and self-learning systems? How can I protect data collections against external attacks, or against intentional and unintentional manipulation? Encryption and privacy-enhancing technologies are just two possibilities: The basic design of trustworthy data spaces and mechanisms for data provisioning by design, i.e. from the outset, are other essential factors.

The international Gaia-X initiative, in which the DLR Institute for AI Safety and Security is involved in a number of areas, is one of the places where current research on security-critical data is taking place. From mobility to Industry 4.0, from culture to health, the numerous Gaia-X domains deal with data security and at the same time with sovereignty, transparency and data availability - the basis for data-based value creation according to European principles. In this way, secure, decentralised and networked data spaces can be created, enabling the emergence of new business models with digital AI-based products. The flexibility of the data spaces allows a wide variety of organisations to develop and operate robust AI applications.