TheoLab-Project Algorithmic Futures

Large language models are increasingly becoming part of the interpretative infrastructures through which historical, theological and cultural knowledge is accessed and processed. Yet different models do not necessarily interpret the same material in the same way. Algorithmic Futures examines how LLMs translate historical and theological positions into future scenarios and which epistemic, normative and interpretative patterns become visible in these projections.

The project develops comparative experimental methods for the study of generative AI in theology and the humanities. Its central methodological approach is futurization: models are confronted with historical source material and asked to imagine what a future world might look like if the position articulated in the source became dominant. By systematically varying models, prompts, languages and contextual information, the project investigates how different LLMs reproduce, transform, criticize or reframe historical positions. The resulting outputs are analysed through a structured annotation framework and made available as research data for comparative study.

Algorithmic Futures thereby treats large language models not merely as tools for retrieving or generating information, but as objects of empirical and epistemological investigation. It asks how model choice, prompting and training-related differences shape the interpretation of historical and theological material and which forms of normative ordering emerge in machine-generated projections. The project thus contributes to Computational Theology by developing methods for the critical and empirically grounded study of generative AI as an emerging infrastructure of theological and humanistic interpretation.