Large Language Models (LLMs) increasingly shape global discourse, making fairness and ideological neutrality essential for responsible AI deployment. Despite growing attention to political bias in LLMs, prior work largely focuses on high-resource, Western languages or narrow multilingual settings, leaving cross-lingual consistency and safe post-hoc mitigation underexplored. To address this gap, we present a large-scale multilingual evaluation of political bias spanning 50 countries and 33 languages. We introduce a complementary post-hoc mitigation framework, Cross-Lingual Alignment Steering (CLAS), designed to augment existing steering methods by aligning ideological representations across languages and dynamically regulating intervention strength. This method aligns latent ideological representations induced by political prompts into a shared ideological subspace, ensuring cross lingual consistency, with the adaptive mechanism prevents over correction and preserves coherence. Experiments demonstrate substantial bias reduction along both economic and social axes with minimal degradation in response quality. The proposed framework establishes a scalable and interpretable paradigm for fairness-aware multilingual LLM governance, balancing ideological neutrality with linguistic and cultural diversity.
El Movimiento de Izquierda Revolucionaria fue un episodio político que condicionó el marco político venezolano. Un marco político en construcción. Fruto de la efervescencia revolucionaria de los años 60 en América Latina, este partido joven, izquierdista y revolucionario fue una expresión del momento histórico que le tocó experimentar. En este artículo se analiza no solo el surgimiento del partido, sino también sus posicionamientos ideológicos, sus relaciones con otros actores y las frustraciones heredadas de la etapa guerrillera.
1789-, Labor in politics. Political activity of the working class
YouTube has emerged as a major platform for political communication and news dissemination, particularly during high-stakes electoral periods. In the context of the 2024 European Parliament and French legislative elections, this study investigates how political actors and news media used YouTube to shape public discourse. We analyze over 100,000 video transcripts and metadata from 74 French YouTube channels operated by national news outlets, local media, and political figures. To identify the key themes emphasized during the campaign period, we applied a semi-automated method that combined large language models with clustering and manual review. The results reveal distinct thematic patterns across the political spectrum and media types, with right-leaning news outlets focusing on topics like immigration, while left-leaning emphasized protest and media freedom. Themes generating the most audience engagement, measured by comment-to-view ratios, were most often the most polarizing ones. In contrast, less polarizing themes such as video games and nature showed higher approval, reflected in like-to-view ratios. We also observed a general tendency across all media types to portray political figures in neutral or critical terms rather than favorable ones.
Giovanna Brunna da Silva Justino, Natália Rejane Salim, Regiane Teixeira Silveira
et al.
Trata-se de um estudo qualitativo que teve como objetivo implementar ações de educação em saúde voltadas para a sexualidade da mulher no climatério/menopausa na Estratégia Saúde da Família, assim como buscar a sensibilização de profissionais de saúde para a temática. Foi realizado um grupo com mulheres no climatério/menopausa, guiado por meio da perspectiva da Educação Popular em Saúde, além de um matriciamento com a equipe de enfermagem. Utilizou-se a análise temática para interpretar os dados produzidos. Evidenciou-se a invisibilidade dessa temática do campo da saúde das mulheres nas políticas públicas e nos serviços de saúde, ressaltando a necessidade da criação de espaços de escuta, acolhimento e troca de vivências das diversas transformações desse período, na Atenção Primária à Saúde, visando a garantia dos direitos sexuais e reprodutivos e o atendimento integral à saúde das mulheres no climatério e na menopausa.
Social Sciences, Labor in politics. Political activity of the working class
Igor Teodoro Guimarães, Antônio Augusto Oliveira Gonçalves
O presente artigo examina as políticas de reforma agrária e assistência social no Brasil a partir de uma pesquisa de campo realizada no assentamento fundiário Pastorinhas. A origem desse território está relacionada à promessa de implementação de uma política pública de reforma agrária nos anos 1990, fato esse que levou um contingente de trabalhadoras/es rurais a ocuparem uma fazenda na zona rural de Brumadinho/MG. A pesquisa qualitativa buscou compreender como o acesso à terra e aos programas sociais permitiram que as famílias hoje assentadas no Pastorinhas superassem os desafios enfrentados e construíssem uma vida mais digna.
Social Sciences, Labor in politics. Political activity of the working class
Evandro de Godoi, Márcia Adriana Rosmann, Leonardo Matheus Pagani Benvenutti
O presente relato comunica experiências constituídas nos debates que ocorreram no “Fórum de estudos: leituras de Paulo Freire”, um evento itinerante que acontece no Rio Grande do Sul. Com periodicidade anual, ele se divide em círculos temáticos nos quais são discutidas questões pertinentes à educação das classes populares. Neste relato, discute-se a docência e os desafios dela, além das possibilidades existentes a partir de um viés popular. É preciso sensibilizarmo-nos de que a docência deve ser discutida de maneira crítica e radical a partir dos diferentes lugares ocupados pelos sujeitos. A universidade pode ser um desses locais, no momento em que efetiva a indissociabilidade entre ensino, pesquisa e extensão, de forma a criar espaços para que o diálogo, com as demandas, sobretudo das classes populares, aconteça, e que outros projetos de sociabilidade e formação humana sejam gestados.
Social Sciences, Labor in politics. Political activity of the working class
The increasing growth of social media provides us with an instant opportunity to be informed of the opinions of a large number of politically active individuals in real-time. We can get an overall idea of the ideologies of these individuals on governmental issues by analyzing the social media texts. Nowadays, different kinds of news websites and popular social media such as Facebook, YouTube, Instagram, etc. are the most popular means of communication for the mass population. So the political perception of the users toward different parties in the country is reflected in the data collected from these social sites. In this work, we have extracted three types of features, such as the stylometric feature, the word-embedding feature, and the TF-IDF feature. Traditional machine learning classifiers and deep learning models are employed to identify political ideology from the text. We have compared our methodology with the research work in different languages. Among them, the word embedding feature with LSTM outperforms all other models with 88.28% accuracy.
This study examines the political bias of chatbots powered by large language models, namely ChatGPT and Gemini, in the context of the 2024 European Parliament elections. The research focused on the evaluation of political parties represented in the European Parliament across 27 EU Member States by these generative artificial intelligence (AI) systems. The methodology involved daily data collection through standardized prompts on both platforms. The results revealed a stark contrast: while Gemini mostly refused to answer political questions, ChatGPT provided consistent ratings. The analysis showed a significant bias in ChatGPT in favor of left-wing and centrist parties, with the highest ratings for the Greens/European Free Alliance. In contrast, right-wing parties, particularly the Identity and Democracy group, received the lowest ratings. The study identified key factors influencing the ratings, including attitudes toward European integration and perceptions of democratic values. The findings highlight the need for a critical approach to information provided by generative AI systems in a political context and call for more transparency and regulation in this area.
A quarter of US adults regularly get their news from YouTube. Yet, despite the massive political content available on the platform, to date no classifier has been proposed to identify the political leaning of YouTube videos. To fill this gap, we propose a novel classifier based on Bert -- a language model from Google -- to classify YouTube videos merely based on their titles into six categories, namely: Far Left, Left, Center, Anti-Woke, Right, and Far Right. We used a public dataset of 10 million YouTube video titles (under various categories) to train and validate the proposed classifier. We compare the classifier against several alternatives that we trained on the same dataset, revealing that our classifier achieves the highest accuracy (75%) and the highest F1 score (77%). To further validate the classification performance, we collect videos from YouTube channels of numerous prominent news agencies, such as Fox News and New York Times, which have widely known political leanings, and apply our classifier to their video titles. For the vast majority of cases, the predicted political leaning matches that of the news agency.
Thanasis Troboukis, Kelly Kiki, Antonis Galanopoulos
et al.
This chapter introduces a research project titled "Analyzing the Political Discourse: A Collaboration Between Humans and Artificial Intelligence", which was initiated in preparation for Greece's 2023 general elections. The project focused on the analysis of political leaders' campaign speeches, employing Artificial Intelligence (AI), in conjunction with an interdisciplinary team comprising journalists, a political scientist, and data scientists. The chapter delves into various aspects of political discourse analysis, including sentiment analysis, polarization, populism, topic detection, and Named Entities Recognition (NER). This experimental study investigates the capabilities of large language model (LLMs), and in particular OpenAI's ChatGPT, for analyzing political speech, evaluates its strengths and weaknesses, and highlights the essential role of human oversight in using AI in journalism projects and potentially other societal sectors. The project stands as an innovative example of human-AI collaboration (known also as "hybrid intelligence") within the realm of digital humanities, offering valuable insights for future initiatives.
Independent fact-checking organizations have emerged as the crusaders to debunk fake news. However, they may not always remain neutral, as they can be selective in the false news they choose to expose and in how they present the information. They can deviate from neutrality by being selective in what false news they debunk and how the information is presented. Prompting the now popular large language model, GPT-3.5, with journalistic frameworks, we establish a longitudinal measure (2018-2023) for political neutrality that looks beyond the left-right spectrum. Specified on a range of -1 to 1 (with zero being absolute neutrality), we establish the extent of negative portrayal of political entities that makes a difference in the readers' perception in the USA and India. Here, we observe an average score of -0.17 and -0.24 in the USA and India, respectively. The findings indicate how seemingly objective fact-checking can still carry distorted political views, indirectly and subtly impacting the perception of consumers of the news.
Codebooks -- documents that operationalize concepts and outline annotation procedures -- are used almost universally by social scientists when coding political texts. To code these texts automatically, researchers are increasing turning to generative large language models (LLMs). However, there is limited empirical evidence on whether "off-the-shelf" LLMs faithfully follow real-world codebook operationalizations and measure complex political constructs with sufficient accuracy. To address this, we gather and curate three real-world political science codebooks -- covering protest events, political violence and manifestos -- along with their unstructured texts and human labels. We also propose a five-stage framework for codebook-LLM measurement: preparing a codebook for both humans and LLMs, testing LLMs' basic capabilities on a codebook, evaluating zero-shot measurement accuracy (i.e. off-the-shelf performance), analyzing errors, and further (parameter-efficient) supervised training of LLMs. We provide an empirical demonstration of this framework using our three codebook datasets and several pretrained 7-12 billion open-weight LLMs. We find current open-weight LLMs have limitations in following codebooks zero-shot, but that supervised instruction tuning can substantially improve performance. Rather than suggesting the "best" LLM, our contribution lies in our codebook datasets, evaluation framework, and guidance for applied researchers who wish to implement their own codebook-LLM measurement projects.
Predicting roll call votes through modeling political actors has emerged as a focus in quantitative political science and computer science. Widely used embedding-based methods generate vectors for legislators from diverse data sets to predict legislative behaviors. However, these methods often contend with challenges such as the need for manually predefined features, reliance on extensive training data, and a lack of interpretability. Achieving more interpretable predictions under flexible conditions remains an unresolved issue. This paper introduces the Political Actor Agent (PAA), a novel agent-based framework that utilizes Large Language Models to overcome these limitations. By employing role-playing architectures and simulating legislative system, PAA provides a scalable and interpretable paradigm for predicting roll-call votes. Our approach not only enhances the accuracy of predictions but also offers multi-view, human-understandable decision reasoning, providing new insights into political actor behaviors. We conducted comprehensive experiments using voting records from the 117-118th U.S. House of Representatives, validating the superior performance and interpretability of PAA. This study not only demonstrates PAA's effectiveness but also its potential in political science research.
AbstractFrom the end of the 1960s until the outbreak of the Civil War (1975), Lebanon experienced a phase of relatively sustained industrial expansion. Albeit the “boom” did not modify significantly Lebanon's tertiarized economic structure, it was anyway sufficient to create the structural conditions for the emergence of a new militant working-class able to become one of the most relevant contentious actors of its time. This new working class was made primarily of very young and recently urbanized unemployed of rural origin, brutally injected in a crude and hyper-exploitative productive cycle where formal labor unions were, for the most part, absent or scarcely effective. The input for their grassroots, transgressive organization into factory-based Workers’ Committees came from the Organization for Communist Action in Lebanon (OACL), i.e. the most important force of the so-called Lebanese New Left, within the framework of a broader process of militant penetration of the “revolutionary classes” produced by the contradictions of Lebanese capitalism. This created the precondition for the Committees to affirm themselves not only as the radical avant-garde of the Lebanese labor movement but also as an integral part of a broader process of contestation of the existing status quo by the subaltern groups emerged from - or activated by - the structural and cultural changes that the country was experiencing. By retrieving the forgotten history of the Workers’ Committees, the article wants to examine the forms and the trajectories whereby such a new working class became an integral part of this process. In particular, by adopting a Gramscian methodology, the article will first expose the structural changes in the Lebanese industrial sector in the examined period and their labor implications. Then, it will focus on the dynamics which superseded the Committees' birth and affirmation, reserving particular attention to the role played by the OACL. Finally, it will conclude by examining the impact of their agency on the political developments that the country was experiencing. The paper contends that the emergence and the affirmation of counter-hegemonical and transformative working-class activism on the eve of the Civil War, along with representing a direct by-product of structural stresses and constraints, was significantly debtor also of the new ideological and militant infrastructures that the emergence of an Arab New Left had contributed to popularize and deploy. The paper wants also to intervene in the historiographical debate on the Lebanese Civil War, stressing the importance of both subaltern actors and class phenomena in its outbreak, which have generally been widely disregarded by the dominant understandings of the conflict.
Recent research has demonstrated that the multi-task fine-tuning of multi-modal Large Language Models (LLMs) using an assortment of annotated downstream vision-language datasets significantly enhances their performance. Yet, during this process, a side effect, which we termed as the "multi-modal alignment tax", surfaces. This side effect negatively impacts the model's ability to format responses appropriately -- for instance, its "politeness" -- due to the overly succinct and unformatted nature of raw annotations, resulting in reduced human preference. In this paper, we introduce Polite Flamingo, a multi-modal response rewriter that transforms raw annotations into a more appealing, "polite" format. Polite Flamingo is trained to reconstruct high-quality responses from their automatically distorted counterparts and is subsequently applied to a vast array of vision-language datasets for response rewriting. After rigorous filtering, we generate the PF-1M dataset and further validate its value by fine-tuning a multi-modal LLM with it. Combined with novel methodologies including U-shaped multi-stage tuning and multi-turn augmentation, the resulting model, Clever Flamingo, demonstrates its advantages in both multi-modal understanding and response politeness according to automated and human evaluations.
Sérgio Trombetta, Jaime José Zitkoski, Valter Marciano dos Santos Chereta
A discussão central que apresentamos neste artigo trata do desafio da pesquisa como horizonte político-pedagógico da educação contemporânea. Diante de um contexto social que nos apresenta um mundo em permanente transformação, a educação deve transmitir e oportunizar o acesso cada vez mais amplo aos saberes que potencializam a leitura crítica do mundo em seus diferentes aspectos. Ler e interpretar o contexto, que muda rapidamente, requer um novo perfil na docência e estratégias mobilizadoras dos educandos para se sentirem estimulados no paradigma do aprender a aprender, ou seja, aprender a pesquisar. Isso porque aprendemos a vida toda e, nesse desafio, cada vez mais o direito à vida vai se confundir com o direito de aprender. Vida e conhecimento estão implicados mutuamente. A metodologia deste estudo se restringe a uma pesquisa bibliográfica e, entre as principais conclusões, podemos destacar que a perspectiva da pesquisa como paradigma pedagógico requer o redimensionamento da vida e das rotinas escolares, para que todos os sujeitos sejam estimulados e mobilizados à construção de uma sociedade aprendente e comprometida eticamente com o futuro sustentável da vida em nosso planeta, com respeito ao pluralismo cultural e à dignidade da pessoa humana.
Social Sciences, Labor in politics. Political activity of the working class
Lidiane Alves de Deus, Marco Antonio Moreira de Oliveira, Rodrigo Moreira Braz
et al.
The presence of women in the educational sphere and in the labor market, in addition to being a struggle with historical roots, has contemporary characteristics rooted in society, requiring knowledge and diagnosis of the most diverse forms of female performance, here dealing with the public environment. Therefore, the objective of this article was to carry out a temporal survey on the representation of women in the teaching position within the coordination of extension activities in a public university, with study object being the Viçosa Federal University, Rio Paranaíba Campus, State of Minas Gerais, Brazil. Data collection was carried out on websites that provide reports on servers and on the institution's internal system for recording outreach activities. Among the surveys with greater evidence and theoretical proof, it could be diagnosed that women, although they do not represent the majority of university professors, are predominantly more present in the coordination of different types of extension activities (programs, projects, courses and events) than men, with emphasis on actions related to the Institute of Biological and Health Sciences, which have greater representation.
Social Sciences, Labor in politics. Political activity of the working class
Social Media have been extensively used for commercial and political communication, besides their initial scope of providing an easy-to-use outlet to produce and consume user-generated content. Besides being a popular medium, Social Media have definitely changed the way we express ourselves or where we look for emerging news and commentary, especially during troubled times. In this paper, we examine a corpus assembled from the Twitter accounts of politicians in the United States and annotated with respect to their audience and the sentiment they convey with each post. Our purpose is to examine whether there are stylistic differences among representatives of different political ideologies, directed to different audiences or with dissimilar agendas. Our findings verify existing knowledge from conventional written communication and can be used to evaluate the quality and depth of political expression and dialogue, especially during the period leading to an election.
Human Activity Recognition (HAR) on mobile devices has been demonstrated to be possible using neural models trained on data collected from the device's inertial measurement units. These models have used Convolutional Neural Networks (CNNs), Long Short-Term Memory (LSTMs), Transformers or a combination of these to achieve state-of-the-art results with real-time performance. However, these approaches have not been extensively evaluated in real-world situations where the input data may be different from the training data. This paper highlights the issue of data heterogeneity in machine learning applications and how it can hinder their deployment in pervasive settings. To address this problem, we propose and publicly release the code of two sensor-wise Transformer architectures called HART and MobileHART for Human Activity Recognition Transformer. Our experiments on several publicly available datasets show that these HART architectures outperform previous architectures with fewer floating point operations and parameters than conventional Transformers. The results also show they are more robust to changes in mobile position or device brand and hence better suited for the heterogeneous environments encountered in real-life settings. Finally, the source code has been made publicly available.