Uses WildChat
WildClaims: Conversational Information Access in the Wild(Chat)
ECIR 2026 · First arXiv submission
Examines factual information conveyed by chatbot responses, including when the user's main task is editing or creative writing. It releases extracted claims and check-worthiness annotations to study information access beyond explicit questions.
Data used
3,000 sampled English WildChat conversations after excluding math and coding.
Findings
WildClaims contains 121,905 extracted claims from 7,587 assistant utterances; estimated check-worthy-conversation prevalence varies substantially with the annotation method.
Section 3 and Table 2: selection and annotation counts. The prevalence analysis compares classification methods. Read the source.
Dates and versions
First arXiv version; the ECIR 2026 paper uses the revised title shown here. ArXiv version history.
ECIR 2026 proceedings paper under its revised title; first posted as a preprint in September 2025.
Cite this paper
Hideaki Joko, Shakiba Amirshahi, Charles L. A. Clarke, Faegheh Hasibi. 2026. WildClaims: Conversational Information Access in the Wild(Chat). Advances in Information Retrieval. ECIR 2026.
Show BibTeX
@inproceedings{joko2026wildclaims,
title = {{WildClaims: Conversational Information Access in the Wild(Chat)}},
author = {Joko, Hideaki and Amirshahi, Shakiba and Clarke, Charles L. A. and Hasibi, Faegheh},
year = {2026},
booktitle = {Advances in Information Retrieval. ECIR 2026},
publisher = {Springer, Cham},
volume = {16486},
pages = {367--382},
doi = {10.1007/978-3-032-21321-1_47},
url = {https://link.springer.com/chapter/10.1007/978-3-032-21321-1_47}
}