Uses WildChat
Adopt ≠ Adapt: Longitudinal Analyses of LLM Conversations in the Wild
arXiv preprint · First arXiv submission
Compares individual and population-level usage patterns in Bing Copilot and WildChat-4.8M. The study tracks changes over a user’s history and compares users with different activity levels across the two platforms.
Data used
WildChat-4.8M, with hashed-IP filters and a main-analysis cutoff before September 2024 to reduce later API-like activity.
Findings
WildChat users’ intent and domain distributions more closely resemble high-activity Copilot users than low-activity users.
Section 3.2: hashed-IP filters and time cutoff. Section 6 and Figure 6: activity groups across the two sources. Read the source.
Dates and versions
First arXiv version; annotation follows the August 2026 revision. ArXiv version history.
2026 arXiv preprint, August revision.
Cite this paper
Rebecca M. M. Hicke, Kiran Tomlinson. 2026. Adopt ≠ Adapt: Longitudinal Analyses of LLM Conversations in the Wild. arXiv preprint.
Show BibTeX
@misc{hicke2026adoptadapt,
title = {{Adopt ≠ Adapt: Longitudinal Analyses of LLM Conversations in the Wild}},
author = {Hicke, Rebecca M. M. and Tomlinson, Kiran},
year = {2026},
doi = {10.48550/arXiv.2605.29018},
eprint = {2605.29018},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2605.29018}
}