Springe direkt zu Inhalt

Delineating the Role of Chatbot Personas as Design Objects in Conversational Agent Design

Requirements

  • a basic understanding of the Human-Centered Design Process will help in writing the thesis
  • Python
Academic Advisor
Discipline
Conversational agent design
Degree
Bachelor of Science (B.Sc.)

Contents

Context

Conversational User Interfaces can limit interface design, but also introduce novel design aspects. The chatbot personality plays a central role in conversational agent design. It is found to support users in building a consistent picture of the chatbot as an object that they can interact with and influences the chatbot's acceptance by users [1][2].

In the following, we refer to the chatbot personality as the combination of perceived humanoid aspects which are interpreted as traits and/or actions by the users. In contrast, a chatbot persona is understood as a design object in conversational agent design which fixes one intended personality interpretation of the chatbot and thereby enables the chatbot development team to interact with and communicate about central design aspects. Having set a chatbot persona may support consistent subsequent design decisions, e.g., during interaction and - more specific - dialog design [3].

In literature on text- and speech-based conversational agents, a number of design dimensions concerning the "personality" of chatbots are studied, e.g. design categories of embodied conversational agents (e.g., appearance/looks, textual output) [4], interpreted personality traits of speech-based conversational agents (e.g., confrontational, polite) [1].

In the BMBF-funded research project DocTalk, the overall goal is to design and implement a pro-active and transparent conversational agent that sustainably supports assistant physicians in their communication and learning processes in the context of their daily clinical routine. Future studies in DocTalk aim at studying the effect of the chatbot personality on its ability to support sustainable learning (i.e., reflective learning). In the research project DocTalk, participatory design methods are applied, which involve users as domain experts who are offered opportunities to actively participate in the design process. By enabling users to interact with design objects (e.g., prototyping toolboxes, prototypes) throughout the process, researchers offer opportunities for users to introduce their domain knowledge to the design process.

Problem

Research literature on structured, participatory approaches to designing personalities of text-based conversational agents is sparse, while best practices abound in the business world [3][5][6].

In this context, this bachelor thesis takes on the challenge of organizing the existing research(!) findings on text-based conversational agent personas. These insights will then be applied to the DocTalk research project in participatory design approach to chatbot persona development. Here, the focus lies on identifying general requirements for chatbot personas from research literature and designing and implementing an initial chatbot persona for a chatbot in the medical context that supports learning processes.

Objectives

  • Identify existing research literature on chatbot personalities and the chatbot persona as a design object in conversational agent design
  • Contextualize chatbot persona design in the iterative human-centered design process of conversational agents (= In how far is there a need for chatbot personas in conversational agent design?)
  • Design and implement a prototypical chatbot persona in a Human-Centered Design Process, including the following aspects:
    • persona as a textual design object
    • image as a visual representation

Possible Procedure

  • Literature review on chatbot persona (context: (medical) education, research field: HCI)
  • Collect data on users and context (user study with residents or advanced medical trainees)
  • Design a prototypical chatbot persona
  • Evaluate the prototypical chatbot persona in a qualitative user study (optional) or else in a heuristic evaluation
  • Discuss the prototypical chatbot persona

References

[1] Völkel, Sarah Theres, u. a. „Developing a Personality Model for Speech-based Conversational Agents Using the Psycholexical Approach“. Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems, Association for Computing Machinery, 2020, S. 1–14. ACM Digital Library, https://doi.org/10.1145/3313831.3376210.

[2] Chen, Zhifa, u. a. „Creating a Chatbot for and with Migrants: Chatbot Personality Drives Co-Design Activities“. Proceedings of the 2020 ACM Designing Interactive Systems Conference, ACM, 2020, S. 219–30. DOI.org (Crossref), doi:10.1145/3357236.3395495.

[3] Google Developers. „Create a persona“. Google Assistant: Conversational Design Process, 27. Februar 2021, https://developers.google.com/assistant/conversation-design/create-a-persona.

[4] ter Stal, Silke, u. a. „Design Features of Embodied Conversational Agents in EHealth: A Literature Review“. International Journal of Human-Computer Studies, Bd. 138, Juni 2020, S. 102409. ScienceDirect, doi:10.1016/j.ijhcs.2020.102409.

[5] The Manifest. „How to Give Your Chatbot a Personality“. Medium, 11. Dezember 2018, https://chatbotslife.com/how-to-give-your-chatbot-a-personality-5e0fb239b28c.

[6] Ultan Ó Broin. „Personality Brings Life to Chatbot User Experience“. Medium, 14. Oktober 2017, https://chatbotsmagazine.com/avoiding-a-clash-of-personalities-chatbot-design-is-no-different-3f0bcd30defd.