Artificial intelligence makes computer software validation (CSV) easier, more reliable, less expensive and more efficient. This is the vision of “valid-AI-te”, a new research project of Saarbrücken-based DHC Business Solutions. Cooperation partners are the University of Mannheim and the Heidelberg-based biotech specialist Ticeba GmbH.
Software systems must be validated. Pharmaceutical companies know this; medical device manufacturers know this; and biotechnology companies have to deal with this. But what does it mean? “Computer Software Validation” – or CSV – is used to prove that a software system actually works in practice as it is intended and as it is written in the documentation. There are norms, guidelines and standards for this; it must be possible to prove this and check it during audits – and for companies this can be very time-consuming and costly.
But does it really have to be that way? Are there no “smarter” solutions for this? This is the question that DHC Business Solutions’ new research project is investigating.
“valid-AI-te” explores whether software validation can be automated through the use of Artificial Intelligence, or AI. There are scientific methods and complex procedures of data analysis for this purpose. They record exactly how a software system behaves; they check whether what happens in the actual application actually takes place as it should. And they document in detail what state a software is in.
“If we succeed in using AI to automate validation to a large extent, then in the future it will not be necessary for every function, every process step, every action to have to be manually rechecked and documented”, emphasizes Dr. Wolfgang Kraemer, Managing Director of DHC Business Solutions. “The software virtually ‘validates itself.'” For companies, this reduces regulatory complexity; small and medium-sized companies in particular benefit.”
In the project “valid-AI-te”: GxP-compliant computer software validation through process automation research approaches from the AI fields of process mining, desktop activity mining, process automation, and process prediction will be tested. Funding comes from the German Federal Ministry of Education and Science (Bundesministerium für Bildung und Wissenschaft, BMBF) as part of the program “KI4KMU – Research, Development and Use of Artificial Intelligence Methods in SMEs” (FKZ 01IS21044A). DHC’s project partners are the Junior Professorship for Management Analytics of Professor Dr. Jana-Rebecca Rehse at the University of Mannheim and Ticeba GmbH from Heidelberg, which has been developing stem cell-based drugs for rare diseases for more than 15 years.