@inproceedings{lai-etal-2026-users,
title = "Users Mispredict What Drives Their Preferences for {AI} Writing Assistance",
author = "Lai, Vivian and
Bu{\c{c}}inca, Zana and
Kang, Hyeonsu B and
Houtti, Mo and
Akpinar, Nil-Jana and
Chian, Kevin and
Suh, Namjoon and
Williams, Alex C.",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-main.1622/",
pages = "36866--36888",
ISBN = "979-8-89176-459-0",
abstract = "Proactive AI writing assistants need to predict when users want drafting help, yet evidence about which contextual cues drive these preferences remains limited. Through a factorial vignette study with 50 participants making 750 pairwise comparisons in professional-email scenarios, we find that compositional effort is the strongest measured predictor ($\rho = 0.597$), whereas urgency has a near-zero scenario-level correlation and no reliable effect across inferential analyses. More critically, we observe a striking stated{--}revealed gap between abstract factor rankings and situated choices: they rank urgency first in self-reports despite it being the weakest behavioral driver, a contrast at the extremes even though mean stated urgency and effort ranks are close, i.e., 2.06 vs. 2.18. This misalignment has measurable consequences. Within this controlled professional-email testbed, the behavior-derived rule agrees with 61.3{\%} of the 750 observed pairwise choices, compared with 57.7{\%} for the stated-preference rule. This 3.6-point difference corresponds to a net gain of 27 matches. It also identifies a concrete design risk for proactive email assistance: stated rankings can place weight on cues that do not predict situated choices. Future studies should test whether stated rankings create similar design risks in other writing tasks."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="lai-etal-2026-users">
<titleInfo>
<title>Users Mispredict What Drives Their Preferences for AI Writing Assistance</title>
</titleInfo>
<name type="personal">
<namePart type="given">Vivian</namePart>
<namePart type="family">Lai</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zana</namePart>
<namePart type="family">Buçinca</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hyeonsu</namePart>
<namePart type="given">B</namePart>
<namePart type="family">Kang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mo</namePart>
<namePart type="family">Houtti</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nil-Jana</namePart>
<namePart type="family">Akpinar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kevin</namePart>
<namePart type="family">Chian</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Namjoon</namePart>
<namePart type="family">Suh</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Alex</namePart>
<namePart type="given">C</namePart>
<namePart type="family">Williams</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-459-0</identifier>
</relatedItem>
<abstract>Proactive AI writing assistants need to predict when users want drafting help, yet evidence about which contextual cues drive these preferences remains limited. Through a factorial vignette study with 50 participants making 750 pairwise comparisons in professional-email scenarios, we find that compositional effort is the strongest measured predictor (ρ = 0.597), whereas urgency has a near-zero scenario-level correlation and no reliable effect across inferential analyses. More critically, we observe a striking stated–revealed gap between abstract factor rankings and situated choices: they rank urgency first in self-reports despite it being the weakest behavioral driver, a contrast at the extremes even though mean stated urgency and effort ranks are close, i.e., 2.06 vs. 2.18. This misalignment has measurable consequences. Within this controlled professional-email testbed, the behavior-derived rule agrees with 61.3% of the 750 observed pairwise choices, compared with 57.7% for the stated-preference rule. This 3.6-point difference corresponds to a net gain of 27 matches. It also identifies a concrete design risk for proactive email assistance: stated rankings can place weight on cues that do not predict situated choices. Future studies should test whether stated rankings create similar design risks in other writing tasks.</abstract>
<identifier type="citekey">lai-etal-2026-users</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.1622/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>36866</start>
<end>36888</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Users Mispredict What Drives Their Preferences for AI Writing Assistance
%A Lai, Vivian
%A Buçinca, Zana
%A Kang, Hyeonsu B.
%A Houtti, Mo
%A Akpinar, Nil-Jana
%A Chian, Kevin
%A Suh, Namjoon
%A Williams, Alex C.
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F lai-etal-2026-users
%X Proactive AI writing assistants need to predict when users want drafting help, yet evidence about which contextual cues drive these preferences remains limited. Through a factorial vignette study with 50 participants making 750 pairwise comparisons in professional-email scenarios, we find that compositional effort is the strongest measured predictor (ρ = 0.597), whereas urgency has a near-zero scenario-level correlation and no reliable effect across inferential analyses. More critically, we observe a striking stated–revealed gap between abstract factor rankings and situated choices: they rank urgency first in self-reports despite it being the weakest behavioral driver, a contrast at the extremes even though mean stated urgency and effort ranks are close, i.e., 2.06 vs. 2.18. This misalignment has measurable consequences. Within this controlled professional-email testbed, the behavior-derived rule agrees with 61.3% of the 750 observed pairwise choices, compared with 57.7% for the stated-preference rule. This 3.6-point difference corresponds to a net gain of 27 matches. It also identifies a concrete design risk for proactive email assistance: stated rankings can place weight on cues that do not predict situated choices. Future studies should test whether stated rankings create similar design risks in other writing tasks.
%U https://aclanthology.org/2026.emnlp-main.1622/
%P 36866-36888
Markdown (Informal)
[Users Mispredict What Drives Their Preferences for AI Writing Assistance](https://aclanthology.org/2026.emnlp-main.1622/) (Lai et al., EMNLP 2026)
ACL
- Vivian Lai, Zana Buçinca, Hyeonsu B Kang, Mo Houtti, Nil-Jana Akpinar, Kevin Chian, Namjoon Suh, and Alex C. Williams. 2026. Users Mispredict What Drives Their Preferences for AI Writing Assistance. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 36866–36888, Budapest, Hungary. Association for Computational Linguistics.