This article focuses on hopes and expectations that are discursively linked to computational technologies in the face of the current "crisis of psychiatric diagnostics". On the basis of document analyses, expert interviews as well as laboratory and conference ethnographies, the fiction of an "unprejudiced gaze" is worked out. According to this imagination, the procedures of "artificial intelligence" may let the data - and ultimately the facts - speak for themselves. However, since "data-driven" research is also determined by conceptual and normative decisions, this fiction could obscure epistemic hierarchizations and ontological prioritizations in psychiatric discourse. Against this backdrop, dependencies and selectivities of research should not be denied but made transparent allowing for a controversial debate., Competing Interests: Der Autor gibt an, dass kein Interessenkonflikt besteht., (Thieme. All rights reserved.)