Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
41 tokens/sec
GPT-4o
59 tokens/sec
Gemini 2.5 Pro Pro
41 tokens/sec
o3 Pro
7 tokens/sec
GPT-4.1 Pro
50 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Persona-L has Entered the Chat: Leveraging LLM and Ability-based Framework for Personas of People with Complex Needs (2409.15604v1)

Published 23 Sep 2024 in cs.HC

Abstract: We present Persona-L, a novel approach for creating personas using LLMs and an ability-based framework, specifically designed to improve the representation of users with complex needs. Traditional methods of persona creation often fall short of accurately depicting the dynamic and diverse nature of complex needs, resulting in oversimplified or stereotypical profiles. Persona-L enables users to create and interact with personas through a chat interface. Persona-L was evaluated through interviews with UX designers (N=6), where we examined its effectiveness in reflecting the complexities of lived experiences of people with complex needs. We report our findings that indicate the potential of Persona-L to increase empathy and understanding of complex needs while also revealing the need for transparency of data used in persona creation, the role of the language and tone, and the need to provide a more balanced presentation of abilities with constraints.

Summary of "Persona-L has Entered the Chat: Leveraging LLM and Ability-based Framework for Personas of People with Complex Needs"

The paper introduces "Persona-L," a novel approach in using LLMs and an ability-based framework to develop personas that more accurately represent individuals with complex needs, such as Down syndrome. Traditional persona creation methods often simplify complex user needs, resulting in stereotypical profiles. By integrating LLMs with an ability-centered perspective, Persona-L aims to offer a more nuanced interaction with personas, enhancing the understanding of users' lived experiences.

Methodology and System Design

The authors describe the design and integration of a web-based interface using LLMs to enable dynamic user interactions. The system employs a Retrieval-Augmented Generation (RAG) approach to ground the LLM’s responses in manually curated data from resources related to Down Syndrome. This method aims to ensure the contextual relevance of interactions, reducing LLM hallucinations. Persona-L is structured into three phases: Profile, Theme/Ability, and Interaction, which guide users in customizing personas and engaging with them.

Key Features and Findings

  1. Ability-Based Framework: The framework shifts focus from disabilities to users' abilities, helping to provide a balanced view that includes potential challenges and strengths. This approach aims to enhance empathy while minimizing stereotypes in persona representations.
  2. Interactivity: Persona-L allows users to interact in real-time, providing immediate insights into personas' experiences. This facilitates a deeper and more authentic understanding, which the authors assert could be particularly beneficial for designers and UX practitioners.
  3. Evaluation through User Study: Six experienced UX professionals tested Persona-L, offering insights into the system’s usability and impact. The paper found the tool effective in enhancing access to personas and improving the empathy-building process. Participants highlighted the need for data transparency and balanced trait portrayal to support decision-making.

Implications and Considerations

The research suggests that using LLMs, along with structured thematic contexts and abilities, can significantly enrich the process of persona creation. This method offers practical implications for UX designers by providing an accessible tool for understanding user needs more comprehensively. The paper advises considering a balance in presenting abilities and pain points to further refine personas.

Future Directions

The authors propose several avenues for future exploration:

  • Extended Longitudinal Studies: Understanding long-term integration of Persona-L in design practices.
  • Co-Creation with Users: Involving users and subject matter experts for more in-depth persona validation.
  • Expanding Data Diversification: Addressing potential biases through broader data sets and more diverse persona generation.

Conclusion

By integrating the advanced capabilities of LLMs with an ability-based framework, Persona-L provides a robust tool for creating dynamic and context-rich personas. This research adds to the discourse on persona development in human-computer interaction, emphasizing the potential of AI-driven methods to produce more empathetic and accurate user representations.

User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (8)
  1. Lipeipei Sun (1 paper)
  2. Tianzi Qin (1 paper)
  3. Anran Hu (13 papers)
  4. Jiale Zhang (36 papers)
  5. Shuojia Lin (1 paper)
  6. Jianyan Chen (1 paper)
  7. Mona Ali (1 paper)
  8. Mirjana Prpa (6 papers)
X Twitter Logo Streamline Icon: https://streamlinehq.com

Tweets