The Middle East Cultural Sensitivity Score (MECSS) is introduced, a framework that turns Said's seven Orientalist operations into measurable dimensions, and the term "Said-washing" for a specific failure: a model that disclaims generalization, then reproduces the structure it disclaimed.
Abstract
AI systems now shape how hundreds of millions of people learn about cultures other than their own. When someone asks one of these systems about the Middle East, they do not receive neutral facts. They receive a representation shaped by the frameworks embedded in training data, and that data is overwhelmingly Western and English-language. This paper asks whether that representation is Orientalist in Said's sense: whether it denies agency to Middle Eastern actors, treats Western frameworks as neutral while marking non-Western knowledge as particular, and explains the region through categories it did not produce. Standard fairness metrics cannot answer this, because they detect explicit prejudice rather than structural framing. This paper introduces the Middle East Cultural Sensitivity Score (MECSS), a framework that turns Said's seven Orientalist operations into measurable dimensions, and the term"Said-washing"for a specific failure: a model that disclaims generalization, then reproduces the structure it disclaimed. Across 280 conversations (1,120 exchanges), GPT-4 and Falcon3-7B-Instruct both reproduce Orientalist patterns systematically, through structural positioning rather than open stereotyping. GPT-4 scores moderately (mean MECSS 1.73); Falcon3-7B-Instruct scores higher (2.18), even though it was built in Abu Dhabi and trained with Arabic content. This is evidence against the assumption that building a model regionally makes it less Orientalist, though the models differ in size as well as origin, so geography cannot be isolated as the cause. Epistemic Center, the treatment of Western frameworks as unmarked universals, scores near the top of the scale for both models. Said-washing appears in 87.9% of GPT-4 conversations, a pattern existing metrics cannot see. Reducing this bias requires changing what models learn from, not only adding languages or relocating institutions.
Professional communication is increasingly mediated by LLMs - but do these models serve all users equally? We show that when prompts contain linguistic features more commonly used by women (hedges, tag questions, collective reference), they systematically elicit shorter, less sophisticated, and less formal responses across three document types and four models. These effects persist after controlling for prompt complexity and feature carry-over. Explicit gender cues like sign-off names are encoded in the same representational space as linguistic dialect - suggesting shared underlying mechanisms - yet linguistic register is far more influential, producing large, consistent effects where names produce none. Our results further reveal that post-hoc mitigation is challenging: because these patterns are culturally embedded and outside conscious control, users cannot easily avoid them through strategic self-presentation, and mechanistic analysis reveals that linguistic features are encoded in early transformer layers and entangled with other features. Our work calls for upstream consideration of the influences of linguistic variation to mitigate disparate impacts of LLM-mediated workplace communication.
Comparison of five widely used large language models suggests that AI-generated language may shape how culturally situated perspectives are expressed, with differences across models indicating that AI-generated language may shape how culturally situated perspectives are expressed.
Ashkan Goudarzi, Aylar Naderi Zonouz· Digital Studies in Language...· 0 citations
Benevolence bias is identified and measure, a small but consistent tendency for aligned LLMs to lean toward the kinder, safer, more socially approved answer on value-laden survey questions, and is easy to diagnose and straightforward to fix.
Yuanzi Li, Jun-Hao Wang, Minghui Liu et al.· 0 citations
Measuring political bias in large language models (LLMs) remains challenging as it can manifest through subtle differences in framing, argumentation, and legal reasoning that are difficult to capture with a single metric. In this work, we introduce Poli-Bias, a counterfactual framework for measuring whether LLMs treat legally equivalent conflict scenarios differently depending on the countries involved. Poli-Bias compares responses to paired prompts in which country identities are systematically swapped across diverse geopolitical relationships, legal violations, and reasoning tasks. Rather than reducing bias to a single judgment, our framework decomposes response disparities into five interpretable dimensions, revealing how and where unequal treatment manifests. Across 13 contemporary LLMs spanning diverse model families and sizes, we find that country identities and user affiliations can systematically affect how equivalent actions are described, evaluated, and defended under international law. Our results thus establish Poli-Bias as a fine-grained framework for auditing political even-handedness and sycophancy in LLMs.
Massi-Nissa Abboud, Aladin Djuhera, Elena Cabrio et al.· 0 citations
Middle-range theory, as conceptualised by Merton, is distinguished by its empirical testability and its capacity to bridge abstract theory and empirical social inquiry. In this paper, we ask whether large language models (LLMs) can be used as surrogate human actors in the testing and reformulation of middle-range theories. Focusing on agenda-setting theory, we investigate how agenda priority, issue detailedness and geographical proximity shape public attention and policy attitudes in the context of climate change. Drawing on climate change public discourses on BiliBili, a major Chinese video-sharing social media platform, we design parallel experiments with human participants, LLM-based virtual agents and LLM-based virtual agents embedded with human personas. By comparing the results across three experiments, we explore how LLMs reproduce or diverge from the dynamics anticipated by agenda-setting theory. The study engages with current theoretical debates on the boundaries of LLMs for social theory construction, addressing their potential and limitations not only as tools for simulation but also as platforms of theoretical innovation. Our findings reveal that while LLMs demonstrate high directional fidelity, they struggle with simulating the magnitude and structural complexity of human behaviours. We thus provide a methodological framework for the use of LLMs as heuristic simulators rather than empirical substitutes of human experiments. We argue that the combination of middle-range theory and generative models yields new possibilities for computational social science while also raising fundamental questions about the nature of empirical validation in the age of generative AI.
Pu Yan, Hengyi Li, Nuo Chen et al.· Theory and society· 0 citations
Conversational AI developed by geopolitical rivals reaches citizens worldwide, raising concerns that it could sway public opinion or be rejected as foreign propaganda, with consequences for democratic discourse and information sovereignty. Yet, whether an AI's perceived national origin shapes its persuasive power is unknown. In a preregistered randomized experiment, 403 adults from a nationally representative United States sample held a three-round debate with a chatbot introduced as either American ("DiscoveryAI") or Chinese ("ZhengheAI"), discussing a political or non-political topic. In all conditions, participants actually conversed with the same model (GPT-4o), instructed to argue against their initial position. We combined pre- and post-conversation self-reports of attitudes, trust, and collective narcissism with computational analyses of 1,209 participant turns, including LLM-coded stance and argumentative conduct, stance-sensitive embeddings, and keyword-masked emotion and toxicity classifiers. The conversations produced substantial attitude changes in every condition. Critically, the nationality label affected neither self-reported attitude change nor expressed stance, concessions, counterarguing, or affect, and equivalence tests and Bayes factors largely supported these null effects. The label's only reliable footprint was lower pre-conversation human-like trust in the Chinese model, whereas functionality trust was unaffected. Political topics slowed stance movement toward the AI's position, and collective narcissism predicted less attitude change regardless of origin, acting as a general barrier rather than an out-group filter. Users thus initially withhold social trust from a rival's AI yet still assimilate its arguments; origin labeling and transparency requirements alone may offer weak protection against foreign influence operations conducted through conversational AI.
Ningzhi Liu, Yannic Hinrichs, Jonas R. Kunst· 0 citations
A new method, called CW-Net, translates the reasoning process of an autonomous vehicle’s AI system into understandable concepts that explain its behavior.
