The Role of Automation, Bots and AI in Influencing Knowledge of the Past

Wills-Eve, Ben and Murrieta-Flores, Patricia and Kemp, Sandra and Pereda, Javier (2024) The Role of Automation, Bots and AI in Influencing Knowledge of the Past. PhD thesis, Lancaster University.

Text (2024wills-evephd)
2024wills-evephd.pdf - Published Version
Available under License Creative Commons Attribution.
Download (10MB)

Abstract

This thesis explored the role of algorithms, automated computer programs (‘bots’) and artificial intelligence (AI) in the creation, interpretation and sharing of knowledge of the past online. It aimed to explore how these specific types of computational approaches influence the diversity of historical and cultural narratives that are researched and search through the Web, especially by institutions like museums and art galleries. The roles of the wide range of different stakeholders involved in these processes, from researchers to museum curators to software developers to interested users, were explored to better understand the human interactions often at the heart of automated approaches. The historical context of automation, bots and algorithms on the Web is explored with regard to their roles in influencing the production, interpretation and sharing of knowledge about the past. Three case studies follow to explore specific types of automated approaches in more detail: the sharing of art and artefacts from online gallery and museum collections via automated social media accounts (‘social bots’); the interpretation of historical figures and events by ‘chatbots’ through the Amazon Alexa platform; and the creation and sharing of historical knowledge and interpretation of online museum collections through the AI tool ChatGPT. I argue that these processes and their influence on knowledge of the past can only be properly understood through a theoretical approach that focusses on the complex human and computational interactions inherent in the production and sharing of knowledge online. Automated approaches all offer potential benefits to cultural heritage institutions looking to engage wider audiences with interpretations of the past, but existing biases in the data available and algorithmic methods involved poses significant risks to the historical accuracy and trustworthiness of such generated material. Greater collaboration and regulation are needed to prevent singular, uncritical interpretations dominating narratives of the past.