How a person expresses their state of mind with language represents rich information about themself. This project develops and adapts a method to improve precision of diagnosing Major Depressive Disorder and Generalized Anxiety Disorder (henceforth, depression and anxiety, respectively). The method uses artificial intelligence (AI) based on computationally and data heavy transfomer-based language (foundation) models to analyze open-ended text responses. The open-ended method is referred to as computational language assessments, whereby respondents answer questions with free text responses that are analyzed with AI including Natural Language Processing, Deep Learning and Machine Learning. The computational language assessments were originally developed by Kjell et al. (2019) and are highly novel and innovative. Several studies show competitive and even greater validity and reliability of the language assessments as compared to rating scales (e.g., Kjell et al., 2019). These findings have the potential to improve and transform psychological assessments because rating scales are the dominating tool to assess self-reported variables in psychology, whereas AI methods on language are still underdeveloped and therefore rarely used. This project seeks to further develop and adapt these computational language assessments with a more person-centred approach to improve the diagnostic predictions.