Artificial intelligence (AI) and machine learning are rapidly developing areas of computer science whose applications span from expert systems through natural language processing to image processing in various domains. Many results show a great promise in improving healthcare, increasing the efficiency of production systems, increasing security, etc. However, most research effort results are still largely academic, and uptake of both academic and industrial solutions is marginal in clinical practice. Especially in high-risk areas, where there is a big chance of serious harm to involved parties, the uptake of AI methods is extremely slow.

According to Whitepaper on AI and Coordinated Plan on Artificial Intelligence, one of the priorities of the EU is the development of trustworthy and secure AI methods that would be acceptable as (partial) substitutes for traditional techniques in these areas. We believe that the trustworthiness of AI can only be achieved via direct communication between AI experts and application domain experts. This is a long-term process that leads to the development of a common language and the ability to connect deep insights from the application perspective with insights into the internal workings of AI systems and the ability to validate resulting methods in clinical practice.