Certification of AI-based Systems

  • Contact:

    Ali Sunyaev

  • Project Group:

    Ali Sunyaev, Sebastian Lins

Artificial intelligence (AI) is a topic that is becoming increasingly important. The rapid changes and developments force both researchers and companies to continuously deal with AI in order to stay on the cutting edge of the technology. A core objective of existing efforts is the creation of Trustworthy AI (TAI), because only if we trust AI and its results, its full potential can be exploited. For example, if an AI-based system diagnoses a disease and provides appropriate treatment, but neither the doctor nor the patients trust it, even if it could save lives, AI will not help.

Certification can help to achieve TAI by creating transparency about the AI used. Unfortunately, there are no certifications yet and it remains open what can be tested and assessed on an AI to build trust. Based on years of experience in the certification of cloud services, the research group Critical Information Infrastructures has therefore set itself the goal of holistically researching the certification of AI-based systems.