Research Challenges in HCI

Large language models (LLMs), such as GPT-3.5 and GPT-4, are increasingly used in real-world applications. However, their performance in practical scenarios remains underexplored due to their closed-source nature. This paper evaluates the use of GPT-3.5 and GPT-4 for extracting insights from a text corpus to identify research challenges in Human-Computer Interaction (HCI). We extracted 4,392~research challenges across over 100~topics from the 2023 CHI conference proceedings and visualized them for interactive exploration. Our evaluation demonstrates that combining GPT-3.5 and GPT-4 offers a cost-efficient approach for large-scale qualitative text analysis. This cost-efficiency is crucial for prototyping research ideas and analyzing text corpora from various perspectives, with significant implications for using LLMs to mine insights in both academic and practical contexts.