Objectives
Our project, inspired by the Visigothic queen Baddo, will generate a context-based LLM with secondary sources produced by historians who research one of the excision of Christianism: Arianism. This includes data from both Spanish and abroad historians from the twentieth century to the present. Intra-religious conflict is one of the social processes that has generated vast amounts of historiographic production. Knowledge of these intra-religious dynamics is key to understanding the formation of European culture, diversity, and identity. Historians present this knowledge in papers, monographs, clips, and audios.
Developments of a context-based LLM model could contribute faster to shaping and completing this historiographical data. The social and political circumstances that could condition the historiographic discourse will also be included, as well as an analysis of the lives of these historians.
OBJECTIVE 1
Train and fine-tune a humanistic AI model tailored to the historiography about the Visigothic period: This model will be built upon existing AI techniques and models, and fine-tuned using secondary sources about this period, enabling it to extract insights from various types of historiographical documents, writings, videos, and audios. The main goal is to generate new perspectives and conclusions that contribute to the understanding and teaching of Visigothic history from diverse sources.
OBJECTIVE 2
Enhance the objectivity and accuracy of historical interpretations: The BADDO project aims to evaluate the effectiveness of AI assistance in historical analysis by comparing AI-generated outcomes with those generated by human beings, combining the opinion of experts in the field and learners.
OBJECTIVE 3
Evaluate the ethical implications of applying AI to the humanities field: This objective will design an ethical framework including manual checks of the AI outcomes to mitigate ethical risks and biases. Discursive language and historical content will be manually checked to generate ethical guidelines that can be used for future researchers and learners. The project aims to ensure transparency, accountability, and fairness in AI-driven analyses, thereby fostering trust and confidence in the use of AI in the humanities.
OBJECTIVE 4
Provide an AI-assistant to generate novel historiographical perspectives: An open web service integrating the fine-tuned AI model will be developed to provide complementary resources to Humanities learners and researchers. This service will provide access to new secondary historical sources and introduce a fine-tuned AI assistant for the Visigothic period to generate new historiographical theories.