Reproduce “An HCI-Centric Survey and Taxonomy of Human-Generative-AI Interactions” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Jingyu Shi, Rahul Jain, Hyungjun Doh, Ryo Suzuki, Karthik Ramani

Researchers from Purdue University and the University of Calgary conducted an HCI-centric survey of 291 papers on human-generative AI interactions, developing a comprehensive taxonomy with six dimensions to systematize existing knowledge and guide future design.

Key insights

  • Current human-GenAI interaction systems predominantly use direct control modalities, suggesting a lack of focus on more intuitive interaction designs.
  • Most GenAI systems prioritize visualizing generated results over revealing intermediate processes, limiting users' understanding of the AI's internal workings and intention-action mapping.
  • A critical gap exists in the literature regarding discussions on ethical problems related to GenAI, with only a small fraction of surveyed papers addressing concerns like credit assignment, bias, or inappropriate content generation.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.