Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Towards Intent-based User Interfaces: Charting the Design Space of Intent-AI Interactions Across Task Types

Zijian Ding

TL;DR

The paper tackles how to design intent-based user interfaces that translate high-level user objectives into actionable AI tasks across task types. It proposes a three-phase study framework spanning fixed-scope content curation, atomic creative tasks, and complex interdependent tasks, to map intent-AI interaction patterns. Early results from headline-generation and analogy-generation studies show both the sufficiency of simple one-off intents for constrained tasks and the value of guided iteration for creativity and sensemaking, while also highlighting risks from misalignment and harmful outputs. The ongoing study on exploratory visual data analysis investigates how to support both expert and novice users through scaffolds and sensemaking affordances on large displays. Collectively, the work informs IUIs that foster high-level intent expression for experienced users and structured guidance for novices, advancing practical IUI design for diverse analytics and creative workflows.

Abstract

Technological advances continue to redefine the dynamics of human-machine interactions, particularly in task execution. This proposal responds to the advancements in Generative AI by outlining a research plan that probes intent-AI interaction across a diverse set of tasks: fixed-scope content curation task, atomic creative tasks, and complex and interdependent tasks. This exploration aims to inform and contribute to the development of Intent-based User Interface (IUI). The study is structured in three phases: examining fixed-scope tasks through news headline generation, exploring atomic creative tasks via analogy generation, and delving into complex tasks through exploratory visual data analysis. Future work will focus on improving IUIs to better provide suggestions to encourage experienced users to express broad and exploratory intents, and detailed and structured guidance for novice users to iterate on analysis intents for high quality outputs.

Towards Intent-based User Interfaces: Charting the Design Space of Intent-AI Interactions Across Task Types

TL;DR

The paper tackles how to design intent-based user interfaces that translate high-level user objectives into actionable AI tasks across task types. It proposes a three-phase study framework spanning fixed-scope content curation, atomic creative tasks, and complex interdependent tasks, to map intent-AI interaction patterns. Early results from headline-generation and analogy-generation studies show both the sufficiency of simple one-off intents for constrained tasks and the value of guided iteration for creativity and sensemaking, while also highlighting risks from misalignment and harmful outputs. The ongoing study on exploratory visual data analysis investigates how to support both expert and novice users through scaffolds and sensemaking affordances on large displays. Collectively, the work informs IUIs that foster high-level intent expression for experienced users and structured guidance for novices, advancing practical IUI design for diverse analytics and creative workflows.

Abstract

Technological advances continue to redefine the dynamics of human-machine interactions, particularly in task execution. This proposal responds to the advancements in Generative AI by outlining a research plan that probes intent-AI interaction across a diverse set of tasks: fixed-scope content curation task, atomic creative tasks, and complex and interdependent tasks. This exploration aims to inform and contribute to the development of Intent-based User Interface (IUI). The study is structured in three phases: examining fixed-scope tasks through news headline generation, exploring atomic creative tasks via analogy generation, and delving into complex tasks through exploratory visual data analysis. Future work will focus on improving IUIs to better provide suggestions to encourage experienced users to express broad and exploratory intents, and detailed and structured guidance for novice users to iterate on analysis intents for high quality outputs.
Paper Structure (9 sections, 3 figures)

This paper contains 9 sections, 3 figures.

Table of Contents

  1. Introduction
  2. Related work
  3. Fixed-scope Content Curation Tasks
  4. Atomic Creative Tasks
  5. Complex and Interdependent Tasks
  6. Study Plan and Preliminary Results
  7. Study 1 on News Headline Generation: One-off Intent Expression for Fixed-Scope Content Curation Tasks
  8. Study 2 on Cross-Domain Analogy Generation: Intent Iteration for Atomic Creative Tasks
  9. Study 3 (in progress) on Exploratory Visual Data Analysis: Intent Exploration, Iteration, and Sensemaking for Complex and Interdependent Tasks

Figures (3)

  • Figure 1: Study 1 on fixed-scope content curation tasks - interface for human-AI news headline co-creation for guidance + selection + post-editing condition: (A) news reading panel, (B) perspectives (keywords) selection panel (multiple keywords can be selected), (C) headline selection panel with post-editing capability, and (D) difficulty rating slider.
  • Figure 2: Study 2 on atomic creative tasks - the prompt for the intent of generating cross-domain analogies based on Duncker and Lees' radiation problem dunckerProblemsolving1945.
  • Figure 3: Study 3 on exploratory visual data analysis - the participant explored and iterated on analysis intents in natural language, and organized generated visualization on a large display.