Prototypical Human-AI Collaboration Behaviors from LLM-Assisted Writing in the Wild

TL;DR

This paper analyzes how users interact with large language models during complex writing tasks, identifying common collaboration behaviors and how user intents influence these interactions, with implications for LLM alignment.

Contribution

It introduces Prototypical Human-AI Collaboration Behaviors (PATHs), revealing dominant interaction patterns and their relation to user intents in real-world LLM-assisted writing.

Findings

A small set of PATHs explains most user-LLM interaction variation.

User writing intents significantly influence collaboration behaviors.

Identified behaviors include revising, exploring, questioning, and style adjustment.

Abstract

As large language models (LLMs) are used in complex writing workflows, users engage in multi-turn interactions to steer generations to better fit their needs. Rather than passively accepting output, users actively refine, explore, and co-construct text. We conduct a large-scale analysis of this collaborative behavior for users engaged in writing tasks in the wild with two popular AI assistants, Bing Copilot and WildChat. Our analysis goes beyond simple task classification or satisfaction estimation common in prior work and instead characterizes how users interact with LLMs through the course of a session. We identify prototypical behaviors in how users interact with LLMs in prompts following their original request. We refer to these as Prototypical Human-AI Collaboration Behaviors (PATHs) and find that a small group of PATHs explain a majority of the variation seen in user-LLM interaction. These PATHs span users revising intents, exploring texts, posing questions, adjusting style or injecting new content. Next, we find statistically significant correlations between specific writing intents and PATHs, revealing how users' intents shape their collaboration behaviors. We conclude by discussing the implications of our findings on LLM alignment.