AI, a Method for Everything?

Artificial intelligence (AI) is impacting scientific research and methodology as both, an object of study and a tool for inquiring and eliciting data. Social science makes no exception, but is specifically addressed for its focus on reflexivity, complexity, black boxes. Social science of technology, with Science and Technology Studies in particular, questions of embedded values and materialized assumptions within the technology render the use of AI in research a special challenge. Our study uses a Grounded Theory based mixed method literature review to investigate the manifoldness of AI notions in research, mapping out an overview of applications, cases, and studies between a vast range of disciplines and scopes. We are highlighting the variety and specificity of these cases, refraining from forcing a common AI ontology. Instead, we claim that AI holds a notable capacity to connect and integrate several fields and traditions with each other but argue for a meticulous differentiation in doing research below the surface layer of AI labeling. We illustrate this argument by example, presenting our own method, its biases and onto-semantic foundation and processes in detail. The chapter concludes with a plaidoyer for methodological reflexivity, meticulousness, and audacity.