Given the enormous amount of routine and research data collected in (mental) health care, data reuse is a rapidly growing area with high potential for clinical practice and research. Data reuse is further encouraged by the increasing availability and usability of new technologies and the implementation of the FAIR data principles. Artificial intelligence (AI) and machine learning (ML) technologies can help structure and process the large amount of available data in an efficient and reproducible manner. The FAIR data principles encourage researchers to make their primary data available for reuse, or to look for reusable existing data sets before collecting new. Despite these developments, data reuse can be challenging, as data sets are being used for purposes other than originally intended. The aim of this PhD thesis was to investigate how new technologies such as AI can contribute to the successful reuse of clinical data towards improving (mental) health care practice and research. The five studies presented in this thesis all made use of secondary data, demonstrating the practice of data reuse based on a range of available routinely collected or research data sets from different health care settings in the Netherlands. It was found that despite the growing availability of data, the possibility to link and enrich these data, and the emergence of new (AI) techniques such as supervised text classification, automated speech processing, and data visualization, reusing data remains a complicated and lengthy process. AI can certainly further successful data reuse, for example by reproducibly processing large amounts of unstructured data or by “blindly” analyzing privacy sensitive data using a text classification tool. However, critical elements such as data quality, label quality, sample size, class balance, and the scope and granularity of the data set also play an important role. It is of great importance to judge these elements beforehand, because if they are insufficient or do not match with the secondary research purpose, applying AI will most likely not lead to more efficient or successful research.