Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Building Better AI Agents: A Provocation on the Utilisation of Persona in LLM-based Conversational Agents

Guangzhi Sun, Xiao Zhan, Jose Such

TL;DR

This provocation examines what persona means for conversational agents and whether embedding distinct personas in LLM-based CAs is feasible or desirable. It surveys pre-LLM technical and social research, contrasts reality with aspiration in current LLM-era systems, and outlines three persona needs: participant simulation, domain-specific role playing, and brand representation. The paper highlights challenges such as maintaining long-term persona consistency, evaluating persona integrity, and ensuring domain knowledge supports the persona, as well as ethical concerns like deception and stereotype reinforcement. Its contribution lies in synthesizing existing findings, identifying gaps, and outlining directions for responsible, domain-aware deployment of persona-infused CAs with emphasis on evaluation standards and safeguards that impact practical adoption across sectors.

Abstract

The incorporation of Large Language Models (LLMs) such as the GPT series into diverse sectors including healthcare, education, and finance marks a significant evolution in the field of artificial intelligence (AI). The increasing demand for personalised applications motivated the design of conversational agents (CAs) to possess distinct personas. This paper commences by examining the rationale and implications of imbuing CAs with unique personas, smoothly transitioning into a broader discussion of the personalisation and anthropomorphism of CAs based on LLMs in the LLM era. We delve into the specific applications where the implementation of a persona is not just beneficial but critical for LLM-based CAs. The paper underscores the necessity of a nuanced approach to persona integration, highlighting the potential challenges and ethical dilemmas that may arise. Attention is directed towards the importance of maintaining persona consistency, establishing robust evaluation mechanisms, and ensuring that the persona attributes are effectively complemented by domain-specific knowledge.

Building Better AI Agents: A Provocation on the Utilisation of Persona in LLM-based Conversational Agents

TL;DR

This provocation examines what persona means for conversational agents and whether embedding distinct personas in LLM-based CAs is feasible or desirable. It surveys pre-LLM technical and social research, contrasts reality with aspiration in current LLM-era systems, and outlines three persona needs: participant simulation, domain-specific role playing, and brand representation. The paper highlights challenges such as maintaining long-term persona consistency, evaluating persona integrity, and ensuring domain knowledge supports the persona, as well as ethical concerns like deception and stereotype reinforcement. Its contribution lies in synthesizing existing findings, identifying gaps, and outlining directions for responsible, domain-aware deployment of persona-infused CAs with emphasis on evaluation standards and safeguards that impact practical adoption across sectors.

Abstract

The incorporation of Large Language Models (LLMs) such as the GPT series into diverse sectors including healthcare, education, and finance marks a significant evolution in the field of artificial intelligence (AI). The increasing demand for personalised applications motivated the design of conversational agents (CAs) to possess distinct personas. This paper commences by examining the rationale and implications of imbuing CAs with unique personas, smoothly transitioning into a broader discussion of the personalisation and anthropomorphism of CAs based on LLMs in the LLM era. We delve into the specific applications where the implementation of a persona is not just beneficial but critical for LLM-based CAs. The paper underscores the necessity of a nuanced approach to persona integration, highlighting the potential challenges and ethical dilemmas that may arise. Attention is directed towards the importance of maintaining persona consistency, establishing robust evaluation mechanisms, and ensuring that the persona attributes are effectively complemented by domain-specific knowledge.
Paper Structure (15 sections, 1 figure)

This paper contains 15 sections, 1 figure.

Table of Contents

  1. What Does 'Persona' Mean in the Context of Conversational Agents?
  2. Persona in CAs in pre-LLM era
  3. Technical research.
  4. Social research.
  5. Reality or Aspiration?
  6. Persona needs in LLM-based CAs
  7. Participant Simulation
  8. Role Playing in Specific Domains
  9. Brand Representation
  10. Challenges and Caveats
  11. Consistency Is the Top Priority
  12. Are There Effective Ways to Evaluate Persona and Its Consistency?
  13. More than Persona
  14. Ethical Considerations
  15. Conclusion

Figures (1)

  • Figure 1: A screenshot of a dialogue with GPT-4-0125-preview. This suggests that GPT-4 does not embody a specific persona. However, this conclusion is based on the model's output, which may not fully align with the designers' intentions.