CONFERENCEWhile with the development and the widespread availability and use of chatbots – especially conversational Artificial Intelligence (AI) – in personal and public environments the amount of human-computer interaction in natural language has increased dramatically, the discrepancy between experiencing a conversation between humans and the experience of human-computer interaction has not been solved.
Conversational AI can now perform their tasks at a high level, but the overall efforts to humanise chatbots have left not only practical problems (such as the lack of important characteristics of human conversations, which affects the quality of services they provide) unsolved, but also fundamental questions, e.g., what "humanisation" means, unaddressed. Therefore, this project intends to contribute to the attempts to bring the philosophical theories of human communicative experience closer to the empirical findings gained from data analysis.
An interdisciplinary research team of artificial intelligence researchers, philosophers and sociologists in the HumBots project studies user/chatbot conversation logs combining sentiment analysis, phenomenological method, hermeneutics and approaches within critical discourse studies to develop recommendations for optimising chatbot communication, allowing to improve the user experience in communication with chatbots both in the field of public administration and private services.
While closely following the latest developments in the global chatbot industry, the project’s main empirical research focuses on the Latvian public administration chatbot family managed by the Culture Information Systems Centre.