Austria

Multilingual large language models have gained prominence for their proficiency in processing and generating text across languages. Like their monolingual counterparts, multilingual models are likely to pick up on stereotypes and other social biases during training. In this paper, we study a phenomenon we term &quot;stereotype leakage&quot;, which refers to how training a model multilingually may lead to stereotypes expressed in one language showing up in the models&#39; behavior in another. We propose a measurement framework for stereotype leakage and investigate its effect in English, Russian, Chinese, and Hindi and with GPT-3.5, mT5, and mBERT. Our findings show a noticeable leakage of positive, negative, and nonpolar associations across all languages. We find that GPT-3.5 exhibits the most stereotype leakage of these models, and Hindi is the most susceptible to leakage effects.

ACL 2025

Multilingual Large Language Models Leak Human Stereotypes across Language Boundaries

### Welcome to 2025 Annual Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics

Welcome to the 2025 meeting of the Nations of the Americas Chapter of the Association for Computational Linguistics! I am proud to help organize the first NAACL conference to carry the new name of our organization, one that emphasizes inclusion for all of the Americas. I am also pleased to welcome you to Albuquerque, New Mexico, a state whose unique blend of cultural influences will make for an excellent backdrop for NAACL 2025, especially with this year’s special theme on NLP in a Multicultural World. 
**[Continue reading...](https://drive.google.com/file/d/1jX-qGhqVSZZCIrAnJaz798pflu5Irrdn/view?usp=sharing)**

*- Colin Cherry, Google, NAACL 2025 General Chair* 

[![](https://assets.underline.io/markdown_image/1/image/b087f8a4dc5816d6e1a6514e59c59ac3.png)](https://drive.google.com/file/d/1T96GzPqObXrMln2BMByCSXTSizjTg69P/view?usp=sharing)

You need to log in with the email address you registered with. Access credentials have been sent to your email. 

Please be sure to check your spam and other email folders if you do not see an email confirmation right away.

Please log in to explore this event.

To access NAACL 2025 event page you are required to register. Please follow [**this link**](https://2025.naacl.org/registration/registration/) to register. Access will depend on your registration type.

Please register!

NAACL 2025

Welcome to 2025 Annual Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics

Multilingual large language models have gained prominence for their proficiency in processing and generating text across languages. Like their monolingual counterparts, multilingual models are likely to pick up on stereotypes and other social biases during training. In this paper, we study a phenomenon we term "stereotype leakage", which refers to how training a model multilingually may lead to stereotypes expressed in one language showing up in the models' behavior in another. We propose a measurement framework for stereotype leakage and investigate its effect in English, Russian, Chinese, and Hindi and with GPT-3.5, mT5, and mBERT. Our findings show a noticeable leakage of positive, negative, and nonpolar associations across all languages. We find that GPT-3.5 exhibits the most stereotype leakage of these models, and Hindi is the most susceptible to leakage effects.

workshop paper

### Welcome to The 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025)

Message from the General Chair: 
*It is my great pleasure and honor to welcome you to the 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025), held in beautiful Vienna, Austria, from July 27 to August 1, 2025. ACL2025continues our field’s tradition of excellence in scholarship, innovation, and inclusivity, and I am deeply grateful to the many volunteers who have worked tirelessly to bring this event to life.* 
[Read more](https://drive.google.com/file/d/1GI_hvOpjswAuYdUTromfeDiPpCcqidwg/view?usp=sharing)

To access this event page, you need to log in with the **email address you registered with**. Access credentials will be sent to your email from Underline - subject line "Welcome to ACL 2025". Please be sure to check your spam email folder if you do not see an email confirmation right away.

Please log in

To access this event page, you need to log in with the **email address you registered with**. 

Welcome to The 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025)

While the scientific consensus on anthropogenic climate change (CC) is undisputed now for a long time, public discourse is still divided. Considering the case of Europe, in the majority of countries, an influential right-wing party propagates climate scepticism or outright denial. Our work addresses the German party, which represents the second-largest faction in the federal parliament. In order to make the partys discourse on CC accessible to NLP-based analyses, we are compiling the, a collection of parliamentary speeches and other material from various sources. We report on first analyses of this new dataset using sentiment and emotion analysis as well as classification of populist language, which demonstrate clear differences to the language use of the two largest competing parties (social democrats and conservatives). We make the corpus available to enable further studies of the party's rhetoric on CC topics.

AfD-CCC: Analyzing the Climate Change Discourse of a German Right-wing Political Party

Sustainability reports are key instruments for communicating corporate impact, but their unstructured format and varied content pose challenges for large-scale analysis. This paper presents an unsupervised method to annotate paragraphs from sustainability reports against both the Global Reporting Initiative (GRI) and Sustainable Development Goals (SDG) standards. The approach combines structured metadata from GRI content indexes, official GRI–SDG mappings, and text semantic similarity models to produce weakly supervised annotations at scale. To evaluate the quality of these annotations, we train a multi-label classifier on the automatically labeled data and evaluate it on the trusted OSDG Community Dataset. The results show that our method yields meaningful labels and improves classification performance when combined with human-annotated data. Although preliminary, this work offers a foundation for scalable sustainability analysis and opens future directions toward assessing the credibility and depth of corporate sustainability claims.

Unsupervised Sustainability Report Labeling based on the integration of the GRI and SDG standards

Moral judgments in LLMs increasingly capture the attention of researchers in AI ethics domain. This study explores moral judgments of three open-source large language models (LLMs)—Qwen-1.5-14B, Llama3-8B, and DeepSeek-R1 in plausible moral dilemmas, examining their sensitivity to social exposure and collaborative decision-making. Using a dual-process framework grounded in deontology and utilitarianism, we evaluate LLMs' responses to moral dilemmas under varying social contexts. Results reveal that all models are significantly influenced by moral norms rather than consequences, with DeepSeek-R1 exhibiting a stronger action tendency compared to Qwen-1.5-14B and Llama3-8B, which show higher inaction preferences. Social exposure and collaboration impact LLMs differently: Qwen-1.5-14B becomes less aligned with moral norms under observation, while DeepSeek-R1's action tendency is moderated by social collaboration. These findings highlight the nuanced moral reasoning capabilities of LLMs and their varying sensitivity to social cues, providing insights into the ethical alignment of AI systems in socially embedded contexts.

What Counts Underlying LLMs' Moral Dilemma Judgments?

TikTok has emerged as a key platform for discussing polarizing topics, including climate change. Despite its growing influence, there is limited research exploring how content features shape emotional alignment between video creators and audience comments, as well as their impact on user engagement. Using a combination of pretrained and fine-tuned textual and visual models, we analyzed 7,110 TikTok videos related to climate change, focusing on content features such as semantic clustering of video transcriptions, visual elements, tonal shifts, and detected emotions. (1) Our findings reveal that positive emotions and videos featuring factual content or vivid environmental visuals exhibit stronger emotional alignment. Furthermore, emotional intensity and tonal coherence in video speech are significant predictors of higher engagement levels, offering new insights into the dynamics of climate change communication on social media. (2) Our preference learning analysis reveals that comment emotions play a dominant role in predicting video shareability, with both positive and negative emotional responses acting as key drivers of content diffusion. We conclude that user engagement—particularly emotional discourse in comments—significantly shapes climate change content shareability.

Insights into Climate Change Narratives: Emotional Alignment and Engagement Analysis on TikTok

Out-of-context and misattributed imagery is the leading form of media manipulation in today's misinformation and disinformation landscape. The existing methods attempting to detect this practice often only consider whether the semantics of the imagery corresponds to the text narrative, missing manipulation so long as the depicted objects or scenes somewhat correspond to the narrative at hand. To tackle this, we introduce News Media Provenance Dataset, a dataset of news articles with provenance-tagged images. We formulate two tasks on this dataset, location of origin relevance (LOR) and date and time of origin relevance (DTOR), and present baseline results on six large language models (LLMs). We identify that, while the zero-shot performance on LOR is promising, the performance on DTOR hinders, leaving room for specialized architectures and future work.

Dataset of News Articles with Provenance Metadata for Media Relevance Assessment

Understanding how humans perceive and describe food is essential for NLP applications such as semantic search, recommendation, and structured food communication. However, textual similarity often fails to reflect perceptual similarity, which is shaped by sensory experience, wine knowledge, and individual context. To address this, we introduce Sensory Anchors—structured reference points that align textual and perceptual representations. Using Red Wine as a case study, we collect free-form descriptions, metaphor-style responses, and perceptual similarity rankings from participants with varying levels of wine knowledge. These rankings reflect holistic perceptual judgments, with wine knowledge emerging as a key factor. Participants with higher wine knowledge produced more consistent rankings and moderately aligned descriptions, while those with lower knowledge showed greater variability. These findings suggest that structured descriptions based on higher wine knowledge may not generalize across users, underscoring the importance of modeling perceptual diversity. We also find that metaphor-style prompts enhance alignment between language and perception, particularly for less knowledgeable participants. Sensory Anchors thus provide a flexible foundation for capturing perceptual variability in food language, supporting the development of more inclusive and interpretable NLP systems.

Bridging Perceptual Gaps in Food NLP: A Structured Approach Using Sensory Anchors

Phishing attacks represent a significant cybersecurity threat, necessitating adaptive detection techniques. This study explores few-shot Adaptive Linguistic Prompting (ALP) in detecting phishing webpages through the multimodal capabilities of state-of-the-art large language models (LLMs) such as GPT-4o and Gemini 1.5 Pro. ALP is a structured semantic reasoning method that guides LLMs to analyze textual deception by breaking down linguistic patterns, detecting urgency cues, and identifying manipulative diction commonly found in phishing content. By integrating textual, visual, and URL-based analysis, we propose a unified model capable of identifying sophisticated phishing attempts. Our experiments demonstrate that ALP significantly enhances phishing detection accuracy by guiding LLMs through structured reasoning and contextual analysis. The findings highlight the potential of ALP-integrated multimodal LLMs to advance phishing detection frameworks, achieving an F1-score of 0.93—surpassing traditional approaches. These results establish a foundation for more robust, interpretable, and adaptive linguistic-based phishing detection systems using LLMs.

Adaptive Linguistic Prompting (ALP) Enhances Phishing Webpage Detection in Multimodal Large Language Models

The rise of AI-driven manipulation poses significant risks to societal trust and democratic processes. Yet, studying these effects in real-world settings is ethically and logistically impractical, highlighting a need for simulation tools that can model these dynamics in controlled settings to enable experimentation. As human interactions increasingly bleed into the large, sometimes opaque, digital realms of social media and the internet, aspects of misinformation also become increasingly difficult to control for, gauge, and study. To address this, we present a simulation environment designed to explore the impact of manipulation on societal-scale interactions and phenomena with a focus on both the real and the digital worlds. By leveraging large language model (LLM)-based multi-agent simulations, we extend the Concordia framework to include several key innovations: (1) integrating a Mastodon server to enable agents to replicate complex social media interactions, (2) an adapter that uses real-world user data to create more grounded agents and social media content; and (3) multi-modal capabilities that enable our agents to interact using both text and images---just as humans do on social media. This hybrid system allows us to model not only public personas in normal agent-to-agent interactions, but also personas and interactions in the context of social media, with the possibility of examining how the interplay influence system outcomes. We aim for more granular and realistic simulations of social scenarios, such as elections, where robustness against manipulation is critical. Our framework supports diverse agent personas, their generation, and tools to analyze belief shifts and interactions in response to disinformation along with other framework features which we are continuing to build upon. We also provide measurement and analysis tools that track simulation dynamics, evaluate metrics to compare settings, and reveal changes to individual and population-level beliefs.

SandboxSocial: A Sandbox for Social Media Using Multimodal AI Agents

What ethical concerns, if any, do researchers have? We introduce EthiCon, a corpus of 1,580 ethical concern statements extracted from scientific papers published in the ACL Anthology. We extract ethical concern keywords from the statements and show promising results in automating the concern identification process. Through a survey (N=200), we compare the ethical concerns of the corpus to the concerns listed by laypeople and professionals in the field. Finally, we compare our retrieved ethical concerns with existing taxonomies and guidelines, pointing to gaps and actionable insights.

Ethical Concern Identification in NLP: A Corpus of ACL Anthology Ethics Statements

Natural disasters such as tropical cyclones cause annual devastation and take a heavy so- cial cost, as disadvantaged communities are typ- ically hit hardest. Among these communities are the speakers of minority and low-resource languages, who may not be sufficiently in- formed about incoming weather events to pre- pare. This work presents an analysis of the current state of machine translation for natural disasters in the languages of communities that are threatened by them. Results suggest that commercial systems are promising, and that in-genre fine-tuning data are beneficial.

Downloads

Next from ACL 2025

AfD-CCC: Analyzing the Climate Change Discourse of a German Right-wing Political Party

Similar lecture

Does Liking Yellow Imply Driving a School Bus? Semantic Leakage in Language Models

.css-70qvj9{display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}Downloads

Next from ACL 2025

AfD-CCC: Analyzing the Climate Change Discourse of a German Right-wing Political Party

Similar lecture

Does Liking Yellow Imply Driving a School Bus? Semantic Leakage in Language Models

Downloads