top of page
Responsible AI Art.png

Voices of the Past
AI and the Diverse Faces of Medieval England

 

   

 


Project Aims and Objectives


 

  1. Studying Biases in Historical Representations through AI

    • Investigate the prevalence of biases using generative AI-based technologies, such as language (e.g. ChatGPT) and image-based (e.g. DALL.E) tools and techniques. Biases may lead to untruthful representation of people, places and objects.

    • Examine the potential dangers of misrepresenting the Middle Ages, as certain medieval figures or events are sometimes appropriated and are not a truthful representation of the past.
       

  2. Building a Responsible and Representative approaches to developing AI Models pertaining to the sources from the Middle Ages) Recreating the Experiences of Underrepresented Groups

    • Develop an AI model that accurately represents the experiences of people focusing on their ethnicity, belief, faith, gender and the social status of individuals living in England during the 12th and 13th centuries, using the Anglo-Norman corpus (initially).

    • Utilize historical sources, archaeological findings, and other relevant data to inform the model's knowledge base and interactions.
       

  3. Training the AI Model to Acquire a Critical View of Biased European Sources

    • Train the AI model to recognize and critically analyze European sources that dehumanize or can impact upon one’s human rights in a contemporary context.

    • Incorporate a mechanism for the AI to provide context and counternarratives when encountering biased or discriminatory content.
       

  4. Creating an Immersive and Interactive Experience

    • Develop a chatbots and  avatar-based AI system that allows users to interact with virtual representations of individuals from the Middle Ages.

    • Utilize archaeological findings, texts, illuminations, and archival documents to create historically accurate and visually compelling experiences.

    • Incorporate virtual and 3D representations of objects, such as relics or physical artifacts, that tell stories about the Middle Ages and the individuals associated with them.

    • Provide users with a more inclusive, accurate, and broader view of history that goes beyond the concept of "historical figures." 

    • We would encourage practice-based explorations, using experimental Digital Humanities approaches, which could be performances, speculative futures or alternative ways of exhibiting findings.















                                             






                               


                                             Project Outcomes and Impact
       

  • Challenge existing biases and misconceptions about the Middle Ages perpetuated by current AI systems.

  • Provide a platform for underrepresented voices and experiences from the past to be heard and understood.

  • Begin with a focus on England during the 12th and 13th centuries, with the aspiration to expand the project to other times and places in the future, creating a scalable model for responsible AI and representation in historical contexts.

  • Engage the public in a more inclusive and accurate understanding of history through interactive AI experiences.

  • Encourage explorations that can inform creative responsible AI practices.

  • Foster collaboration between academic institutions, museums, and technology partners to create innovative and impactful research projects.

  • Encourage multidisciplinary and experimental research approaches to understanding the interplay of technology, society and research.


    If you are interested, please contact Delphine Demelas and Alan Chamberlain

Project Overview

 

This research project aims to address and mitigate biases in AI representations of historical figures, particularly focusing on the Middle Ages. By critically examining the historical representations perpetuated by AI systems and developing an inclusive approach to developing AI models, we aim to create a more representative and accurate depiction of the Middle Ages. This project will explore the experiences of people focusing on ethnicity, belief, faith, gender and the social status of individuals living in 12th and 13th century England, offering an inclusive perspective through innovative AI technologies.

Charlemagne and Ganelon.jpg

Charlemagne and Ganelon (Paris, BnF, Français 2608 f.147v)
source - 
https://portail.biblissima.fr/

The Queen of Sheba (Göttigen, Staatsbibliothek, Cod. Ms. philos. 63, 122r)
source - BlackCentrealEurope.com

Generative AI (2024) using CoPilot Designer

Let’s Work Together

Get in touch so we can start working together - Join the Network 
 

Dr Alan Chamberlain 

RAiUK_Logo.png
Unknown_edited.png

The R3 Network is supported by the following Engineering and Physical Sciences Research Council projects.

AI UK: Creating an International Ecosystem for Responsible AI Research and Innovation (RAI UK), (RAKE Responsible Innovation Advantage in Knowledge Exchange) [grant number EP/Y009800/1]

Horizon: Trusted Data-Driven Products [grant number EP/T022493/1]
UKRI Trustworthy Autonomous Systems Hub (The TAS RRI II project) [grant number EP/V00784X/1] 

bottom of page