# Unlocking Adaptive User Experience with Generative AI

**Authors:** Yutan Huang, Tanjila Kanij, Anuradha Madugalla, Shruti Mahajan, Chetan Arora, John Grundy  
**arXiv:** [2404.05442](https://arxiv.org/abs/2404.05442)  
**Date:** 2024-04-08  
**Summary page:** https://www.summarizepaper.com/en/paper/2404.05442v1/

## Key points

- Thorough user research is essential in user-centered application development but can be time-consuming and costly.
- Generative AI techniques based on Large Language Models (LLMs) offer potential for creating adaptive interfaces.
- The paper introduces a novel approach using ChatGPT to develop user personas and adaptive interface candidates within a specific domain.
- The study compared outcomes of using ChatGPT with traditional manual processes for developing user personas and adaptive interfaces.
- Data for the personas was collected from 37 survey participants and 4 interviews with a not-for-profit organization.
- Results show promising potential for incorporating LLMs in designing adaptive interfaces.
- Future research aims to explore other aspects of generative AI in UI/UX design and establish a framework for automating persona and website development.

## Summary

In the realm of user-centered application development, conducting thorough user research to cater to diverse user needs is essential but can be time-consuming and costly. However, with the emergence of generative AI techniques based on Large Language Models (LLMs), there is a potential for these powerful tools to be leveraged in creating adaptive interfaces. This paper introduces a novel approach that utilizes ChatGPT to develop user personas and adaptive interface candidates within a specific domain. The study involved comparing the outcomes of developing user personas and adaptive interfaces using both ChatGPT and a traditional manual process. Data for the personas was gathered from 37 survey participants and 4 interviews conducted in collaboration with a not-for-profit organization. requires thorough user research to cater to diverse needs. This can be time-consuming and costly. However,  based on  offer potential for creating . This paper presents an innovative method that uses ChatGPT to develop  and adaptive interface candidates within a specific domain. The study compares outcomes from ChatGPT-generated content with manually created content. Data for the personas was collected through surveys from 37 participants and 4 interviews with a not-for-profit organization. Results show promising potential for incorporating LLMs in designing adaptive interfaces. Future research aims to explore other aspects of generative AI in UI/UX design and establish a robust framework for automating persona and website development. The goal is to capture crucial user-centric content that assists designers in streamlining the design process while enabling fine-tuning and personalization of interactive elements in UI/UX designs. This study provides valuable insights into how  can enhance .