AI and Historical Document Tagging
Chase Carstensen, Jhih-ci Chao, Noah Solomon, Stephen Stardellis
Post Bellum is an organization that preserves and collects historical testimonies and stories to educate the public. They have collected over 9000 interviews as part of their Memory of Nations. Our team used artificial intelligent (AI) to create geographic labels for all these interviews in their database. This would allow users to search the database by location. We determined that Google’s Gemini would be appropriate for this project. We developed a procedure that would label the interviews for past countries, present countries, and Czech regions. Using this method we were able to label the interviews much faster than manually classifying them. We also developed a computer program that would scan the internet for potential colleagues who would be interested in a newsletter from Post Bellum.
See the final report here.