Interaction-Required Suggestions for Control, Ownership, and Awareness in Human-AI Co-Writing

TL;DR

The paper argues that current large language model interfaces excessively favor turn-taking and output generation, limiting human involvement in writing tasks. It introduces interaction-required designs that foster cognitive engagement, granular control, and visibility of the solution space, demonstrated through two techniques: predictive-text typing to compose the assistant's response and highlight-based visualization of alternative edits. A case study on revision illustrates how these interactions promote thoughtful decision-making and ownership, while enabling finer-grained guidance and exploration of possibilities. The work suggests that such interaction designs can augment human-AI co-writing by making the collaboration more deliberate, transparent, and controllable, with future work focusing on empirical evaluation and richer visualization of alternatives.

Abstract

This paper explores interaction designs for generative AI interfaces that necessitate human involvement throughout the generation process. We argue that such interfaces can promote cognitive engagement, agency, and thoughtful decision-making. Through a case study in text revision, we present and analyze two interaction techniques: (1) using a predictive-text interaction to type the assistant's response to a revision request, and (2) highlighting potential edit opportunities in a document. Our implementations demonstrate how these approaches reveal the landscape of writing possibilities and enable fine-grained control. We discuss implications for human-AI writing partnerships and future interaction design directions.

Figures (2)

• Figure 1: Predictive text interaction repurposed to type the assistant's response
• Figure 2: Highlighting opportunities for divergent choices