Chatbot Research and Design: 7th International Workshop, CONVERSATIONS 2023, Oslo, Norway, November 2223, 2023, Revised Selected Papers (Lecture Notes in Computer Science, 14524)

This book constitutes the proceedings of the 7th International Workshop on Chatbot Research and Design, CONVERSATIONS 2023, which was held during November 2023.The 12 regular papers were carefully reviewed and selected for inclusion in the book. They were organized in following topical sections: Understanding and Enhancing Conversational Interactions, LLM-driven Conversational Design and Analysis, Ethical Perspectives and Bias, Complementing Perspectives. Preface Organization Contents Understanding and Enhancing Conversational Interactions Voting Assistant Chatbot for Increasing Voter Turnout at Local Elections: An Exploratory Study 1 Introduction 2 Theory and Hypotheses 2.1 The Use of Communication Means in Municipal Elections 2.2 The Role of Cognitive Processes in the Use of Voting Aids 2.3 Conversational Agent Voting Advice Application 2.4 The Effect of CAVAA on Political Knowledge and Voting Intention 2.5 The Impact of CAVAA on Young People 3 Method 3.1 Research Design 3.2 Participants 3.3 Procedure 3.4 The Conversational Agent Voting Advice Application Design 3.5 Measuring Instrument 3.6 Data Analysis 4 Results 4.1 The Effect of the Voting Assistant Chatbot on Perceived Political Knowledge 4.2 The Effect of the Voting Assistant Chatbot on the Voting Intention of Potential Voters 4.3 The Effect of the Voting Assistant Chatbot on Young Adults and the Elderly 4.4 The Effect of the Voting Assistant Chatbot on Lower, Secondary and Higher Educated Voters 5 Discussion References Conversational Repair Strategies to Cope with Errors and Breakdowns in Customer Service Chatbot Conversations 1 Introduction 2 Theoretical Framework 3 Study 1 3.1 Method 3.2 Results and Conclusion 4 Study 2 4.1 Method 4.2 Results 5 Discussion 6 Conclusion References Aragón Open Data Assistant, Lesson Learned of an Intelligent Assistant for Open Data Access 1 Introduction 2 Technological Context 2.1 Aragón Open Data 2.2 Ontology and Knowledge Graphs Datastore 2.3 Linguistic Models 3 Description of the System 3.1 System Process Flow 3.2 Architecture 3.3 Intention Classification and General Answering Module 4 User Testing Methodology 4.1 Pre-assessment and User Selection 4.2 Approach User Tests and Methodology to Follow 4.3 Results of the First Iteration of User Tests 4.4 Results of the Second Iteration of User Tests 4.5 Improvements Proposed After User Tests 5 Conclusion and Lessons Learned References LLM-driven Conversational Design and Analysis Saleshat: A LLM-Based Social Robot for Human-Like Sales Conversations 1 Introduction 2 Related Work 2.1 Social Robots in Retail 2.2 Large Language Models 3 Design and Development of Saleshat 3.1 Objectives and Requirements 3.2 System Architecture 4 Evaluation 5 Discussion 5.1 Contributions and Implications 5.2 Limitations and Future Work 6 Conclusion References Leveraging Large Language Models as Simulated Users for Initial, Low-Cost Evaluations of Designed Conversations 1 Introduction 2 Related Work 2.1 Evaluating Conversational Agents 2.2 Simulated Users 2.3 Tilbot and Prompt Engineering 3 Preliminary Evaluations 3.1 Autoethnographic Case Studies 3.2 Small-Scale Evaluation 4 Discussion and Conclusion References Examining Lexical Alignment in Human-Agent Conversations with GPT-3.5 and GPT-4 Models 1 Introduction 2 Related Work 3 Method 3.1 Research Design 3.2 Data Collection 3.3 Data Preprocessing 3.4 Data Analysis 4 Results 4.1 Role of GPT Model Version 4.2 Role of Conversation Topic 5 Discussion 5.1 Major Findings 5.2 Limitations and Future Work 6 Conclusion References Ethical Perspectives and Bias In Search of Dark Patterns in Chatbots 1 Introduction 2 Related Work 2.1 Background 2.2 Dark Patterns and Conversational User Interfaces 3 Methodology 3.1 ChIPS Dataset 3.2 Inclusion/Exclusion Criteria 4 Results 5 Discussion 6 Limitations and Threats to Validity 7 Conclusion and Future Work References Language Ideology Bias in Conversational Technology 1 Introduction 2 Language Ideologies, Policies and Public Discourse 3 Method 4 Results and Discussion 4.1 Study I 4.2 Study II 4.3 Discussion and Potential Actions 5 Conclusions References Should Conversational Agents Care About Our Gender Identity? 1 Introduction 2 State of the Art 3 Method 3.1 Design and Participants 3.2 Stimulus Material 3.3 Procedure 3.4 Measurements 3.5 Analyses 4 Results 4.1 Manipulation Check 4.2 Assumptions 4.3 Main Model 5 Discussion 6 Conclusions References Complementing Perspectives Conversational Interactions with NPCs in LLM-Driven Gaming: Guidelines from a Content Analysis of Player Feedback 1 Introduction 2 Related Work 3 Method 3.1 Steam Reviews of Vaudeville 3.2 Analysis 4 Findings 4.1 Flexible and Open-Ended Conversations 4.2 NPC Personality and Conversational Style 4.3 Inappropriate and Unexpected NPC Social Responses 4.4 NPC Hallucinations 4.5 NPC Memory 4.6 Input Modalities 5 Guidelines for Developing LLM-Driven NPCs 6 Discussion 7 Conclusion References Exploring the Dark Corners of Human-Chatbot Interactions: A Literature Review on Conversational Agent Abuse 1 Introduction 2 Addressing Agent Abuse: A Pressing Challenge 2.1 Agent Abuse: A Widespread Phenomenon 2.2 Agent Abuse: A Concerning Phenomenon 3 A Literature Review of Agent Abuse 3.1 Research Questions and Development 3.2 Method and PRISMA Chart 4 Insights 4.1 RQ1: What is the State of the Art of Research Around Conversational Agent Abuse? 4.2 RQ2: From a Counter-Action Perspective: How Can Conversational Agent Abuse Be Limited? 5 Conclusions and Future Paths References Anticipating User Needs: Insights from Design Fiction on Conversational Agents for Computational Thinking 1 Introduction 2 Related Work 3 Research Design 3.1 Research Approach 3.2 Method 4 Results 4.1 Expectations 4.2 Sentiments 5 Discussion 6 Limitations 7 Conclusion References Author Index