Development of a Conversational AI Chatbot Framework for the Arts and Tourism Contents

Roles

• Participated as an UX Research Intern

• User Experience Lab, Seoul National University

Goals

• Defining user’s difficulties and speech patterns when they search arts arts and tourism contents

• Construcing a speech ontology for the AI chatbot assistant

Details

• Funded by Korea Creative Content Agency

• Crawled data from the website dealing with cultural content (e.g. online reviews) and other CAs (Conversational Agents)

• Analyzed the speech patterns of people in searching scenarios

As a research intern at Seoul National University, I had the opportunity to study the difficulties users go through when searching for cultural (e.g., arts, music) content and mitigated the complexity of domain knowledge with an ontology-based AI chatbot assistant. In a conventional setting, cultural contents like historical artifacts are fine-grained into various categories in multiple depths (let alone the terminologies that appear along with them), making it difficult for users to search, explore and learn from them. In response, our team analyzed the speech patterns of people in such scenarios and constructed a speech ontology for the AI chatbot assistant. During the project, I realized how a communicative AI can act as an effective conduit to expert knowledge for people, highlighting the importance of human-AI interaction. Combined with a natural language-powered AI like BERT, these platforms can also teach laypeople to behave like experts while serving them with the necessary knowledge to do so.