@inproceedings{ochieng-etal-2026-toward,
title = "Toward Cultural Alignment: Human-Centered Evaluation of Multimodal {AI} Stories Across Five {A}frican Communities",
author = "Ochieng, Millicent and
Ali, Felermino D. M. A. and
Ankrah, Elizabeth A. and
Abdulhamid, Najeeb Gambo and
Boyd, Migisha and
Nyairo, Stephanie and
Muchai, Mercy and
Maina, Samuel Chege and
Vashistha, Aditya and
Thieme, Anja and
O{'}Neill, Jacki",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.1946/",
doi = "10.18653/2026.findings-emnlp.1946",
pages = "41484--41510",
ISBN = "979-8-89176-462-0",
abstract = "In this paper, we examine how well AI-generated multimodal stories align with the lived practices, relationships, language, values, and visual expectations of the communities they represent. We conduct a community-grounded mixed-methods evaluation with 19 culture representatives across five African communities, combining quantitative annotations with qualitative focus group discussions. We find that cultural alignment depends not simply on recognizable cultural markers, but on how those markers fit social, linguistic, procedural, and visual context. From these evaluations, we develop a taxonomy of cultural alignment comprising five broader cultural marker categories and eight recurring mechanisms of misalignment. We additionally evaluate five multimodal LLM judges to examine whether automated evaluation can approximate community-grounded judgments at scale. Judge reliability and score calibration vary substantially across communities, with no single judge performing consistently across all five settings. These findings motivate community-calibrated evaluation pipelines in which automated judges are validated against community judgments to determine where they can be trusted and where human review remains necessary."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="ochieng-etal-2026-toward">
<titleInfo>
<title>Toward Cultural Alignment: Human-Centered Evaluation of Multimodal AI Stories Across Five African Communities</title>
</titleInfo>
<name type="personal">
<namePart type="given">Millicent</namePart>
<namePart type="family">Ochieng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Felermino</namePart>
<namePart type="given">D</namePart>
<namePart type="given">M</namePart>
<namePart type="given">A</namePart>
<namePart type="family">Ali</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Elizabeth</namePart>
<namePart type="given">A</namePart>
<namePart type="family">Ankrah</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Najeeb</namePart>
<namePart type="given">Gambo</namePart>
<namePart type="family">Abdulhamid</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Migisha</namePart>
<namePart type="family">Boyd</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Stephanie</namePart>
<namePart type="family">Nyairo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mercy</namePart>
<namePart type="family">Muchai</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Samuel</namePart>
<namePart type="given">Chege</namePart>
<namePart type="family">Maina</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Aditya</namePart>
<namePart type="family">Vashistha</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Anja</namePart>
<namePart type="family">Thieme</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jacki</namePart>
<namePart type="family">O’Neill</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>Findings of the Association for Computational Linguistics: EMNLP 2026</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">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>In this paper, we examine how well AI-generated multimodal stories align with the lived practices, relationships, language, values, and visual expectations of the communities they represent. We conduct a community-grounded mixed-methods evaluation with 19 culture representatives across five African communities, combining quantitative annotations with qualitative focus group discussions. We find that cultural alignment depends not simply on recognizable cultural markers, but on how those markers fit social, linguistic, procedural, and visual context. From these evaluations, we develop a taxonomy of cultural alignment comprising five broader cultural marker categories and eight recurring mechanisms of misalignment. We additionally evaluate five multimodal LLM judges to examine whether automated evaluation can approximate community-grounded judgments at scale. Judge reliability and score calibration vary substantially across communities, with no single judge performing consistently across all five settings. These findings motivate community-calibrated evaluation pipelines in which automated judges are validated against community judgments to determine where they can be trusted and where human review remains necessary.</abstract>
<identifier type="citekey">ochieng-etal-2026-toward</identifier>
<identifier type="doi">10.18653/2026.findings-emnlp.1946</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.1946/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>41484</start>
<end>41510</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Toward Cultural Alignment: Human-Centered Evaluation of Multimodal AI Stories Across Five African Communities
%A Ochieng, Millicent
%A Ali, Felermino D. M. A.
%A Ankrah, Elizabeth A.
%A Abdulhamid, Najeeb Gambo
%A Boyd, Migisha
%A Nyairo, Stephanie
%A Muchai, Mercy
%A Maina, Samuel Chege
%A Vashistha, Aditya
%A Thieme, Anja
%A O’Neill, Jacki
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F ochieng-etal-2026-toward
%X In this paper, we examine how well AI-generated multimodal stories align with the lived practices, relationships, language, values, and visual expectations of the communities they represent. We conduct a community-grounded mixed-methods evaluation with 19 culture representatives across five African communities, combining quantitative annotations with qualitative focus group discussions. We find that cultural alignment depends not simply on recognizable cultural markers, but on how those markers fit social, linguistic, procedural, and visual context. From these evaluations, we develop a taxonomy of cultural alignment comprising five broader cultural marker categories and eight recurring mechanisms of misalignment. We additionally evaluate five multimodal LLM judges to examine whether automated evaluation can approximate community-grounded judgments at scale. Judge reliability and score calibration vary substantially across communities, with no single judge performing consistently across all five settings. These findings motivate community-calibrated evaluation pipelines in which automated judges are validated against community judgments to determine where they can be trusted and where human review remains necessary.
%R 10.18653/2026.findings-emnlp.1946
%U https://aclanthology.org/2026.findings-emnlp.1946/
%U https://doi.org/10.18653/2026.findings-emnlp.1946
%P 41484-41510
Markdown (Informal)
[Toward Cultural Alignment: Human-Centered Evaluation of Multimodal AI Stories Across Five African Communities](https://aclanthology.org/2026.findings-emnlp.1946/) (Ochieng et al., Findings 2026)
ACL
- Millicent Ochieng, Felermino D. M. A. Ali, Elizabeth A. Ankrah, Najeeb Gambo Abdulhamid, Migisha Boyd, Stephanie Nyairo, Mercy Muchai, Samuel Chege Maina, Aditya Vashistha, Anja Thieme, and Jacki O’Neill. 2026. Toward Cultural Alignment: Human-Centered Evaluation of Multimodal AI Stories Across Five African Communities. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 41484–41510, Budapest, Hungary. Association for Computational Linguistics.