Taxonomy of interaction behaviors with AI code suggestions

Research Method · All research methods
Evidence · 2 claims (1 for, 1 against) · 1 study (1 associational), q2 · 1 of 1 report an effect size · 2 claims rest on one study

Description

The article synthesizes fragmented prior characterizations of AI code completion interactions into a unified taxonomy of behavioral metrics. It organizes events into suggestion generation events (generate attention check, generate suggestion, suggestion shown), acceptance events (tab accept, slow accept), rejection and revision events (accept then modify, accept then delete, ignore), execution events (run code), and derived and aggregated session metrics such as acceptance rate, failed attention check rate, and average dwell time. The taxonomy operationalizes behaviors previously observed only qualitatively, such as slow accept, backtracking, and accept-and-adapt, so they can be captured quantitatively at scale.

Accounts

• Taxonomy of interaction behaviors with AI code suggestions: The article synthesizes fragmented prior characterizations of AI code completion interactions into a unified taxonomy of behavioral metrics. It organizes events into suggestion generation events (generate attention check, generate suggestion, suggestion shown), acceptance events (tab accept, slow accept), rejection and revision events (accept then modify, accept then delete, ignore), execution events (run code), and derived and aggregated session metrics such as acceptance rate, failed attention check rate, and average dwell time. The taxonomy operationalizes behaviors previously observed only qualitatively, such as slow accept, backtracking, and accept-and-adapt, so they can be captured quantitatively at scale. (Jessica Hutchison et al. (2026))

Claims

• Log analysis reveals novel post-acceptance behaviors (slow tab accept, slow accept then modify) that show suggestion acceptance is a multi-step process [+M]
• Behavioral interaction metrics are weakly or inconsistently correlated with task performance in AI-assisted programming [-W]

Related Research Methods

• Four-metric behavioral taxonomy for distinguishing authentic iterative coding from automated workflows

Key Sources

• Jessica Hutchison, Ian Tyler Applebaum, Kenneth Angelikas, Kush Rakesh Patel, Phuoc Nguyen, Antonio Lazaro, Nicholas Rucinski, Rahad Arman Nabid, and Stephen MacNeil. (2026). To Tab or Not to Tab: Measuring Critical Engagement in AI Code Completion Tools Using Behavioral Signals and Attention Checks. Proceedings of the 31st ACM Conference on Innovation and Technology in Computer Science Education (ITiCSE 2026). https://doi.org/10.1145/3803400.3809394

Page metadata

Field	Value
type	research-method
id	taxonomy-of-interaction-behaviors-with-ai-code-suggestions
title	Taxonomy of interaction behaviors with AI code suggestions
description	The article synthesizes fragmented prior characterizations of AI code completion interactions into a unified taxonomy of behavioral metrics.
status	review
generated	by process:settle-candidates on 2026-10-10
trust	unverified — no one has confirmed this page's content

Frontmatter is the page's machine-readable record — see the Schema & Guide.