One of the project ideas I proposed at the Hackathon was to create a chatbot for art exhibitions. Over the next few days, my team and I built a limited but functional prototype. We called it Albot – The Art History Chatbot.
You’re at a museum and want to find out more about an artwork you like? You no longer have to distract yourself and others, just to take some blurry photos – photos that you’ll struggle to find later anyway. The solution is simple: Just ask Albot, the art history chatbot! Type-in the number or name of the artwork and Albot immediately sends you a high-quality image, along with all the essential information, straight to your phone. As a simple Facebook‑, Twitter- or Text-Message. Even years later, you can search and find every artwork in seconds. But that’s just the beginning. Albot can do much more. He will access the museum’s metadata for you and answer questions about the artwork, like: Who’s the artist? What’s the title? What were the historical circumstances? What does it depict, refer to or mean? What’s the technique? What’s the conservation history? What did people write about it? What was its influence?
There’s no limit to the extent of information, such a Chatbot is able to provide. It doesn’t have to be just plain text either! It can be hyperlinks, images, audio or video as well. But what’s more important than any amount of information. The chatbot initiates a conversation. It’s not a didactic, dogmatic monologue. It’s a dialogue. A dialogue between the viewer, the artwork and art history.
Using Albrecht Dürer’s famous “Allerheiligenbild” as our prototypical exemplar, we collected, analyzed and processed all the information and meta-data that’s available in digital repositories. We conducted extensive user-interviews and collected a wide range of questions that users actually have. Then, the team members who were arthistoric experts about everything related to Dürer, compiled a corpus of answers to these questions. We surveyed and tried out several chatbot frameworks, each had their particular strenghts and significant weaknesses. The night before the final presentation, I’ve implemented a simple but functional chatbot, using the Dexter framework. I learned that the most important part of building a chatbot is the script. It’s all about designing the conversation. To make it fun, playful and engaging. Not just a collection of stiltedly worded factoids. Conversational empathy comes first. The technology is secondary.
My special thanks and gratitude to Dr. Harald Klinke, Head of the Coding Dürer Hackaton for inviting me! Also, thanks to the members of my team and to all the amazing participants of the hackathon!
Dr. Harald Klinke, M.Sc., Assistant Professor, Editor of the International Journal for Digital Art History, Institut für Kunstgeschichte, Ludwig-Maximilians-University München.
[Unfortunately, Albot – The Art History Chatbot is not online anymore.]