Research Papers

Large language models (LLMs) are increasingly shaping citizens' information ecosystems. Products incorporating LLMs, such as chatbots and AI Companions, are now widely used for decision support and information retrieval, including in sensitive domains, raising concerns about hidden biases and growing potential to shape individual decisions and public opinion. This paper introduces GermanPartiesQA, a benchmark of 418 political statements from German Voting Advice Applications across 11 elections to evaluate six commercial LLMs. We evaluate their political alignment based on role-playing experiments with political personas. Our evaluation reveals three specific findings: (1) Factual limitations: LLMs show limited ability to accurately generate factual party positions, particularly for centrist parties. (2) Model-specific ideological alignment: We identify consistent alignment patterns and the degree of political steerability for each model across temperature settings and experiments. (3) Claim of sycophancy: While models adjust to political personas during role-play, we find this reflects persona-based steerability rather than the increasingly popular, yet contested concept of sycophancy. Our study contributes to evaluating the political alignment of closed-source LLMs that are increasingly embedded in electoral decision support tools and AI Companion chatbots.

As an AI Language Model, "Yes I Would Recommend Calling the Police": Norm Inconsistency in LLM Decision-Making

Shomik Jain, D Calacci, Ashia Wilson

We investigate the phenomenon of norm inconsistency: where LLMs apply different norms in similar situations. Specifically, we focus on the high-risk application of deciding whether to call the police in Amazon Ring home surveillance videos. We evaluate the decisions of three state-of-the-art LLMs -- GPT-4, Gemini 1.0, and Claude 3 Sonnet -- in relation to the activities portrayed in the videos, the subjects' skin-tone and gender, and the characteristics of the neighborhoods where the videos were recorded. Our analysis reveals significant norm inconsistencies: (1) a discordance between the recommendation to call the police and the actual presence of criminal activity, and (2) biases influenced by the racial demographics of the neighborhoods. These results highlight the arbitrariness of model decisions in the surveillance context and the limitations of current bias detection and mitigation strategies in normative decision-making.

HyperCLOVA X Technical Report

Kang Min Yoo, Jaegeun Han, Sookyo In +393

We introduce HyperCLOVA X, a family of large language models (LLMs) tailored to the Korean language and culture, along with competitive capabilities in English, math, and coding. HyperCLOVA X was trained on a balanced mix of Korean, English, and code data, followed by instruction-tuning with high-quality human-annotated datasets while abiding by strict safety guidelines reflecting our commitment to responsible AI. The model is evaluated across various benchmarks, including comprehensive reasoning, knowledge, commonsense, factuality, coding, math, chatting, instruction-following, and harmlessness, in both Korean and English. HyperCLOVA X exhibits strong reasoning capabilities in Korean backed by a deep understanding of the language and cultural nuances. Further analysis of the inherent bilingual nature and its extension to multilingualism highlights the model's cross-lingual proficiency and strong generalization ability to untargeted languages, including machine translation between several language pairs and cross-lingual inference tasks. We believe that HyperCLOVA X can provide helpful guidance for regions or countries in developing their sovereign LLMs.

Disentangling and Operationalizing AI Fairness at LinkedIn

Joaquin Quiñonero-Candela, Yuwen Wu, Brian Hsu +5

Operationalizing AI fairness at LinkedIn's scale is challenging not only because there are multiple mutually incompatible definitions of fairness but also because determining what is fair depends on the specifics and context of the product where AI is deployed. Moreover, AI practitioners need clarity on what fairness expectations need to be addressed at the AI level. In this paper, we present the evolving AI fairness framework used at LinkedIn to address these three challenges. The framework disentangles AI fairness by separating out equal treatment and equitable product expectations. Rather than imposing a trade-off between these two commonly opposing interpretations of fairness, the framework provides clear guidelines for operationalizing equal AI treatment complemented with a product equity strategy. This paper focuses on the equal AI treatment component of LinkedIn's AI fairness framework, shares the principles that support it, and illustrates their application through a case study. We hope this paper will encourage other big tech companies to join us in sharing their approach to operationalizing AI fairness at scale, so that together we can keep advancing this constantly evolving field.

Causality-Aided Trade-off Analysis for Machine Learning Fairness

Zhenlan Ji, Pingchuan Ma, Shuai Wang +1

There has been an increasing interest in enhancing the fairness of machine learning (ML). Despite the growing number of fairness-improving methods, we lack a systematic understanding of the trade-offs among factors considered in the ML pipeline when fairness-improving methods are applied. This understanding is essential for developers to make informed decisions regarding the provision of fair ML services. Nonetheless, it is extremely difficult to analyze the trade-offs when there are multiple fairness parameters and other crucial metrics involved, coupled, and even in conflict with one another. This paper uses causality analysis as a principled method for analyzing trade-offs between fairness parameters and other crucial metrics in ML pipelines. To ractically and effectively conduct causality analysis, we propose a set of domain-specific optimizations to facilitate accurate causal discovery and a unified, novel interface for trade-off analysis based on well-established causal inference methods. We conduct a comprehensive empirical study using three real-world datasets on a collection of widelyused fairness-improving techniques. Our study obtains actionable suggestions for users and developers of fair ML. We further demonstrate the versatile usage of our approach in selecting the optimal fairness-improving method, paving the way for more ethical and socially responsible AI technologies.

Gaining the trust of customers and providing them empathy are very critical in the financial domain. Frequent occurrence of fraudulent activities affects these two factors. Hence, financial organizations and banks must take utmost care to mitigate them. Among them, ATM fraudulent transaction is a common problem faced by banks. There following are the critical challenges involved in fraud datasets: the dataset is highly imbalanced, the fraud pattern is changing, etc. Owing to the rarity of fraudulent activities, Fraud detection can be formulated as either a binary classification problem or One class classification (OCC). In this study, we handled these techniques on an ATM transactions dataset collected from India. In binary classification, we investigated the effectiveness of various over-sampling techniques, such as the Synthetic Minority Oversampling Technique (SMOTE) and its variants, Generative Adversarial Networks (GAN), to achieve oversampling. Further, we employed various machine learning techniques viz., Naive Bayes (NB), Logistic Regression (LR), Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF), Gradient Boosting Tree (GBT), Multi-layer perceptron (MLP). GBT outperformed the rest of the models by achieving 0.963 AUC, and DT stands second with 0.958 AUC. DT is the winner if the complexity and interpretability aspects are considered. Among all the oversampling approaches, SMOTE and its variants were observed to perform better. In OCC, IForest attained 0.959 CR, and OCSVM secured second place with 0.947 CR. Further, we incorporated explainable artificial intelligence (XAI) and causal inference (CI) in the fraud detection framework and studied it through various analyses.

Despite the fact that variation is a fundamental characteristic of natural language, automatic speech recognition systems perform systematically worse on non-standardised and marginalised language varieties. In this paper we use the lens of language policy to analyse how current practices in training and testing ASR systems in industry lead to the data bias giving rise to these systematic error differences. We believe that this is a useful perspective for speech and language technology practitioners to understand the origins and harms of algorithmic bias, and how they can mitigate it. We also propose a re-framing of language resources as (public) infrastructure which should not solely be designed for markets, but for, and with meaningful cooperation of, speech communities.

Algorithmic bias mitigation has been one of the most difficult conundrums for the data science community and Machine Learning (ML) experts. Over several years, there have appeared enormous efforts in the field of fairness in ML. Despite the progress toward identifying biases and designing fair algorithms, translating them into the industry remains a major challenge. In this paper, we present the initial results of an industrial open innovation project in the banking sector: we propose a general roadmap for fairness in ML and the implementation of a toolkit called BeFair that helps to identify and mitigate bias. Results show that training a model without explicit constraints may lead to bias exacerbation in the predictions.

Synthetic personae experiments have become a prominent method in Large Language Model alignment research, yet the representativeness and ecological validity of these personae vary considerably between studies. Through a review of 63 peer-reviewed studies published between 2023 and 2025 in leading NLP and AI venues, we reveal a critical gap: task and population of interest are often underspecified in persona-based experiments, despite personalization being fundamentally dependent on these criteria. Our analysis shows substantial differences in user representation, with most studies focusing on limited sociodemographic attributes and only 35% discussing the representativeness of their LLM personae. Based on our findings, we introduce a persona transparency checklist that emphasizes representative sampling, explicit grounding in empirical data, and enhanced ecological validity. Our work provides both a comprehensive assessment of current practices and practical guidelines to improve the rigor and ecological validity of persona-based evaluations in language model alignment research.

We present a novel approach to bias mitigation in large language models (LLMs) by applying steering vectors to modify model activations in forward passes. We compute 8 steering vectors, each corresponding to a different social bias axis, such as age, gender, or race, on a training subset of the BBQ dataset and compare the effectiveness of these to 3 additional bias mitigation methods across 4 datasets. When optimized on the BBQ dataset, our individually tuned steering vectors achieve average improvements of 12.8% on BBQ, 8.3% on CLEAR-Bias, and 1% on StereoSet, and show improvements over prompting and Self-Debias in all cases, and improvements over fine-tuning in 12 out of 17 evaluations. In addition, steering vectors showed the lowest impact on MMLU scores of the four bias mitigation methods tested. The work presents the first systematic investigation of steering vectors for bias mitigation, and we demonstrate that they are a powerful and computationally efficient strategy for reducing bias in LLMs, with broader implications for enhancing AI safety.

Generative AI is transforming education by enabling personalized learning, enhancing administrative efficiency, and fostering creative engagement. This paper explores the opportunities and challenges these tools bring to pedagogy, proposing actionable frameworks to address existing equity gaps. Ethical considerations such as algorithmic bias, data privacy, and AI role in human centric education are emphasized. The findings underscore the need for responsible AI integration that ensures accessibility, equity, and innovation in educational systems.

Universities are increasingly adopting data-driven strategies to enhance student success, with AI applications like Learning Analytics (LA) and Predictive Learning Analytics (PLA) playing a key role in identifying at-risk students, personalising learning, supporting teachers, and guiding educational decision-making. However, concerns are rising about potential harms these systems may pose, such as algorithmic biases leading to unequal support for minority students. While many have explored the need for Responsible AI in LA, existing works often lack practical guidance for how institutions can operationalise these principles. In this paper, we propose a novel Responsible AI framework tailored specifically to LA in Higher Education (HE). We started by mapping 11 established Responsible AI frameworks, including those by leading tech companies, to the context of LA in HE. This led to the identification of seven key principles such as transparency, fairness, and accountability. We then conducted a systematic review of the literature to understand how these principles have been applied in practice. Drawing from these findings, we present a novel framework that offers practical guidance to HE institutions and is designed to evolve with community input, ensuring its relevance as LA systems continue to develop.