AI & Society最新文献

{"title":"Productive power in social networks: challenges for post-phenomenological mediation theory","authors":"João Vidal","doi":"10.1007/s00146-025-02248-3","DOIUrl":"10.1007/s00146-025-02248-3","url":null,"abstract":"<div><p>This article is a philosophical study that examines power and resistance within the conceptual network of post-phenomenological theories of technological mediation. In the contemporary landscape, the interaction between human beings and technology plays a crucial role in shaping social, cultural, and political dynamics. As human beings increasingly emerge into a world permeated by technological innovations, it becomes essential to understand the complex power relations that arise from this intertwining. In this context, post-phenomenological theories of technological mediation offer a valuable theoretical framework capable of exploring the nuances of these interactions. The study begins by positioning power and resistance within this conceptual network. It then explores the complex interactions of power in technologically mediated relationships, with particular emphasis on the technologies incorporated into social media platforms, i.e., algorithmic recommendation systems. Drawing on the concepts of power as defined by Michel Foucault and Byung-Chul Han, the paper challenges conventional understandings of productive power by examining how these platforms shape and exert influence. Ultimately, it introduces the concept of intelligent power as a means to better understand the power dynamics inherent in algorithmic mediation.</p></div>","PeriodicalId":47165,"journal":{"name":"AI & Society","volume":"40 6","pages":"4881 - 4891"},"PeriodicalIF":4.7,"publicationDate":"2025-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s00146-025-02248-3.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144909665","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"WEIRD? Institutions and consumers’ perceptions of artificial intelligence in 31 countries","authors":"Bronwyn Howell","doi":"10.1007/s00146-025-02217-w","DOIUrl":"10.1007/s00146-025-02217-w","url":null,"abstract":"<div><p>A survey of perceptions of Artificial Intelligence in 31 countries in 2023 (Ipsos in Global Views on A.I. 2023. https://www.ipsos.com/sites/default/files/ct/news/documents/2023-07/Ipsos%20Global%20AI%202023%20Report-WEB_0.pdf. Accessed 17 May 2024, 2023) yields significantly less positive perceptions of the new technology in developed western economies than in emerging and non-western economies. This could reflect citizens in non-Western countries perceiving machines (computers) and algorithms differently from those in Western countries, or that a more positive outlook in countries with weak democratic institutions comes from a preference for algorithmic precision over inconsistent and/or corrupt regulation and decision-making. However, it could also be reflecting the different psychology of “WEIRD” (Western, Educated, Industrialised, Rich, Democratic) countries. Regressing the survey responses against measures of the “WEIRD” dimensions, we find that reported understanding of, willingness to trust, and anticipation of change due to AI applications are consistently negatively correlated to a country’s education levels (E), and average income per capita (R). The sophistication of democratic institutions (D) and “Westernness” (W), both alone and in combination with the other factors, have statistically significant negative effects on the percentage of the respondents in any given country having positive perceptions of AI and its prospects. The consistency of the negative relationship between the sophistication of democratic institutions country-level perceptions of AI brings into question the role of regulation of the new technology. WEIRD societies are presumed to rely on democratic institutions for assurances they can transact safely with strangers. Institutions thus substitute for the trust non-WEIRD societies place in friends, family and close community contacts when transacting. Third-party (and notably government) assurances in the context of uncertainty created by the emergence of new AI technologies arguably condition perceptions of the safety of these technologies through the presence (or absence) of regulations governing their implementation and use. Different perceptions amongst European countries compared to other western counterparts to perceptions of data privacy support the contention that the mere presence of AI regulation may be sufficient to alter perceptions in WEIRD societies, regardless of whether the regulations are necessary or even effective in increasing user safety. This has implications for interpreting and responding to political pressure to regulate new technologies in WEIRD countries.</p></div>","PeriodicalId":47165,"journal":{"name":"AI & Society","volume":"40 6","pages":"4409 - 4431"},"PeriodicalIF":4.7,"publicationDate":"2025-02-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s00146-025-02217-w.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144909730","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"“This robot is dictating her next steps in life”: disability justice and relational AI ethics","authors":"Georgia van Toorn,&nbsp;Jackie Leach Scully,&nbsp;Sandra Gendera","doi":"10.1007/s00146-025-02224-x","DOIUrl":"10.1007/s00146-025-02224-x","url":null,"abstract":"<div><p>As automated technologies, particularly artificial intelligence (AI) and automated decision-making (ADM), become integral to social life, there is growing concern about their ethical implications. While issues of accountability, transparency, and fairness dominate discussions on “ethical” AI, little attention has been given to how socially disadvantaged groups most impacted by ADM systems form ethical judgments about them. Drawing on insights from relational ethics, this study uses dialogue groups with disabled people to explore how people distinguish between ‘more just’ or ‘less just’ uses of technology, and the contextual, situational, and relational factors that shape these judgments. For the dialogue group participants in our study, ethical reasoning was most strongly influenced by concerns about how ADM systems affect self-determination, caring relationships and identity recognition, and about the political–economic drivers of automation. The article contributes to AI ethics by empirically demonstrating that justice and ethics depend on the social relationships valued in different contexts and what is at stake, both personally and politically, in decisions aided by automation.</p></div>","PeriodicalId":47165,"journal":{"name":"AI & Society","volume":"40 6","pages":"4473 - 4483"},"PeriodicalIF":4.7,"publicationDate":"2025-02-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s00146-025-02224-x.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144909889","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Algorithmically mediated judgment: an arendtian perspective on political subjectivity in social media","authors":"Anthony Longo","doi":"10.1007/s00146-025-02230-z","DOIUrl":"10.1007/s00146-025-02230-z","url":null,"abstract":"<div><p>This article examines how algorithms mediate the human faculty of judgment within the context of social media. Challenging the common view that algorithms ‘undermine’ or ‘eliminate’ human judgment, I argue instead that they mediate the human-world relations in which judgments emerge. Drawing on Hannah Arendt’s phenomenological approach to identity and judgment, the article deconstructs prevailing assumptions about ‘human judgment’ and ‘algorithmic judgment’ and proposes a different approach to understand user-algorithm relations as co-constitutive rather than oppositional. In line with the ‘empirical turn’ in the philosophy of technology, which emphasizes the importance of grounding philosophical inquiry in lived experience, the proposed philosophical approach is developed and illustrated through a case study of ADHD-related content on TikTok. The case study demonstrates how algorithmic amplification becomes part of self-apprehension and public discourse. Through this lens, the article explores the interplay between self-disclosure, judgment, and the collective formation of political subjectivity, revealing the productive and disruptive roles algorithms play in these processes.</p></div>","PeriodicalId":47165,"journal":{"name":"AI & Society","volume":"40 6","pages":"4905 - 4918"},"PeriodicalIF":4.7,"publicationDate":"2025-02-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144909890","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Large Language Models to make museum archive collections more accessible","authors":"Manon Reusens,&nbsp;Amy Adams,&nbsp;Bart Baesens","doi":"10.1007/s00146-025-02227-8","DOIUrl":"10.1007/s00146-025-02227-8","url":null,"abstract":"<div><p>Keywords are essential to the searchability and therefore discoverability of museum and archival collections in the modern world. Without them, the collection management systems (CMS) and online collections these cultural organisations rely on to record, organise, and make their collections accessible, do not operate efficiently. However, generating these keywords manually is time consuming for these already resource strapped organisations. Artificial intelligence (AI), particularly generative AI and Large Language Models (LLMs), could hold the key to generating, even automating, this key data and as such be considered a co-creative add-on. This study contributes to the literature by introducing the use of Meta’s open-source LLM, Llama, to generate keywords from curator/archivist written descriptions of museum and archival collection items. Our findings suggest that these technologies add significant value compared to current manual methods for keyword generation. In particular, we find that through using carefully crafted prompts, successful keyword augmentations could be established making museum and archival collections much more accessible to wider and more diverse audiences. However, the results also showed that generative AI has biases (e.g., hallucinations, over generalisations, outdated language), though the frequency of occurrence was not as high as general perception may insist. Hence, we also discuss mitigation strategies to address these and how cultural institutions can recognise the risks and errors while getting the most from the systems. Finally, we discuss options to achieve structured results which allow easier ingestion of data back into CMS. Ultimately, LLMs hold significant potential to enhance accessibility to museum and archival collections, yet they are not without imperfection as we extensively discuss.</p></div>","PeriodicalId":47165,"journal":{"name":"AI & Society","volume":"40 6","pages":"4485 - 4497"},"PeriodicalIF":4.7,"publicationDate":"2025-02-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144909729","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Data autonomy and privacy in the smart home: the case for a privacy smart home meta-assistant","authors":"Alexander Orlowski,&nbsp;Wulf Loh","doi":"10.1007/s00146-025-02182-4","DOIUrl":"10.1007/s00146-025-02182-4","url":null,"abstract":"<div><p>In this paper, we focus on privacy risks in smart home environments and their implications for privacy and data protection. As with other Internet of Things (IoT) devices, the collection and processing of user data in smart home environments currently lack transparency and control. Smart home applications operate within the home, a space that is both morally and legally particularly protected and characterized by a implicit expectation of privacy from the user’s perspective. In contrast to these higher privacy risks, the current regulatory efforts are not yet up to speed with respect to smart home environments. As an interim workaround solution, in this paper, we propose a meta-assistant for the smart home that increases users’ data autonomy and thereby their privacy. In the first section, we give a brief overview of smart home applications, their data collection mechanisms, and the implications for user privacy. Following this, we argue in the second section that consent to datafication, i.e., the prevalent legal option to obtain legal grounds for data collection and processing, in most smart home contexts is—albeit legally sufficient—morally inadequate to provide meaningful possibilities for users to exercise their data autonomy and manage their privacy. The third section introduces an interim solution, outlining the possibility of a meta-assistant, which is capable of operating all other devices—if necessary, by shutting them off completely.</p></div>","PeriodicalId":47165,"journal":{"name":"AI & Society","volume":"40 6","pages":"4171 - 4184"},"PeriodicalIF":4.7,"publicationDate":"2025-02-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s00146-025-02182-4.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144909717","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Balancing risks and benefits: public perceptions of AI through traditional surveys and social media analysis","authors":"Daniel Kouloukoui,&nbsp;Nathalie de Marcellis-Warin,&nbsp;Thierry Warin","doi":"10.1007/s00146-025-02232-x","DOIUrl":"10.1007/s00146-025-02232-x","url":null,"abstract":"<div><p>The rapid evolution of Artificial Intelligence (AI) has broadened discussions about its social and technological implications. The purpose of this study is to examine the perception of risks and benefits associated with AI, focusing on privacy, cybersecurity, trust in government management, and online device acceptance over time. We used an integrative approach that combined questionnaires and data analysis from X via Natural Language Processing (NLP). The traditional survey had 1013 participants in 2018 and 1000 in 2021, while X’s analysis examined approximately 7.5 million tweets in the same period. The results indicate a constant concern with the risks of AI, such as the dehumanization of services (86%), job loss (82%), changes in labor functions (89%), privacy (87%), and cyber-attacks (87%). X also highlighted ethical concerns and the use of algorithms, perceived as a “threat cloud.” Acceptance of connected devices was low (23%), reflecting security and privacy concerns. Confidence in government management was remarkably low in relation to cybersecurity (12%) and privacy (13%). Nevertheless, there is a modest increase in the perception of the benefits of AI, such as technological advances and efficiency. This study underscores the need for collaborative and effective policies and the promotion of a culture of responsibility to maximize the benefits of AI while mitigating risks. The practical implications of these findings are discussed, emphasizing the importance of policies and regulations to address emerging digital threats.</p></div>","PeriodicalId":47165,"journal":{"name":"AI & Society","volume":"40 6","pages":"4919 - 4942"},"PeriodicalIF":4.7,"publicationDate":"2025-02-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144909718","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Machine learning methods for isolating indigenous language catalog descriptions","authors":"Yi Liu,&nbsp;Carrie Heitman,&nbsp;Leen-Kiat Soh,&nbsp;Peter Whiteley","doi":"10.1007/s00146-025-02223-y","DOIUrl":"10.1007/s00146-025-02223-y","url":null,"abstract":"<div><p>Museum collection databases contain echoes of encounter between colonial collectors (broadly defined) and Indigenous people from around the world. The moment of acquisition—when an item passed out of a community and into the hands of the collector—often included multilingual acts of translation. An artist may have shared the Indigenous name of the object, or the terms associated with its origin and use. Late nineteenth and twemtieth century museum registrars would in turn transcribe this information from field logs into museum catalogs. Over time, these catalog entries were transformed into digital records within collections managements systems (e.g., EMu, PastPerfect, etc.). As a result of this 150-year process, today’s museum collection databases are riddled with Indigenous words and descriptions, scattered across various metadata fields. They may include Native place-names, family names or vocabulary terms that, when translated, extend far beyond the categories ascribed by museum collection managers. These instances of Indigenous description may also serve as a crucial bridge for reconnecting source communities with items of particular interest to their cultural heritage and linguistic preservation efforts. Aiming to enhance the accessibility of Indigenous languages contained in the metadata of cultural heritage collections, this paper explores applications of machine learning methodologies to identify Indigenous terms present in museum catalogs. Specifically, we discuss methods that incorporate the Google Cloud Language Identification Service to detect A:shiwi (Pueblo of Zuni) language terms through a case study of metadata records from the two largest natural history museums in the USA. We utilize an elimination mechanism to exclude specific languages (e.g., English and Spanish) at the word and phrase levels to detect A:shiwi terms. Our approach outperforms conventional methods in terms of accuracy, recall, precision, and F1-scores. This method can be used to confront the “Digital Heap” of cultural heritage records across institutions to improve the discoverability of Indigenous languages in metadata descriptions and reconnect source communities with items of cultural patrimony.</p></div>","PeriodicalId":47165,"journal":{"name":"AI & Society","volume":"40 6","pages":"4461 - 4471"},"PeriodicalIF":4.7,"publicationDate":"2025-02-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144909709","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Developing computer vision and machine learning strategies to unlock government-created records","authors":"Greg Jansen,&nbsp;Richard Marciano","doi":"10.1007/s00146-025-02231-y","DOIUrl":"10.1007/s00146-025-02231-y","url":null,"abstract":"<div><p>This paper outlines the development of a proof-of-concept workflow using machine learning and computer vision techniques to unlock the data within digitized handwritten US Census forms from the 1950s. The 1950s US Census includes over 6.5 million page images and was only recently made available to the public on April 1, 2022, following a 72-year access restriction period. Our project uses computational treatments to assist researchers in their efforts to recover and preserve the history of the erased Sacramento Japantown. Sacramento once housed the fourth largest Japantown in the United States before experiencing WWII Japanese American Incarceration and the 1950s US Government program of urban renewal. The goal is to augment a researcher’s work in selecting a subset of Census pages for further transcription and analysis. We demonstrate a workflow for extracting demographic information using computer vision for image segmentation, and machine learning for handwritten character recognition. The workflow consists of a computational filtering process for Census records and a user interface for page review. These computational techniques are suitable for other cities, states, and communities, and demonstrate new strategies to unlock vital demographic information. The approach highlights the potential benefits of computational techniques for the analysis of form-based historical records of the twentieth century that can have an impact on social justice.</p></div>","PeriodicalId":47165,"journal":{"name":"AI & Society","volume":"40 6","pages":"4513 - 4529"},"PeriodicalIF":4.7,"publicationDate":"2025-02-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144909676","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"How can we improve the diversity of archival collections with AI? Opportunities, risks, and solutions","authors":"Lise Jaillant,&nbsp;Olivia Mitchell,&nbsp;Eric Ewoh-Opu,&nbsp;Maribel Hidalgo Urbaneja","doi":"10.1007/s00146-025-02222-z","DOIUrl":"10.1007/s00146-025-02222-z","url":null,"abstract":"<div><p>This article is the first study to examine the impact (positive and negative) of Artificial Intelligence on the diversity of archival collections. Representing the diverse audiences they serve is a key objective for libraries and archives. For example, institutions with colonial-era archival documents are experimenting with AI to improve the discoverability of their collections and to enhance access for source communities and other users. Indeed, AI can be used to automatically create metadata, search vast amounts of historical records, and answer questions with natural language. However, these technologies also come with risks—for instance when AI systems are trained on potentially biased data. Very little is known about the impact of these computational tools on diversity in archival collections. Do AI technologies compound or alleviate the lack of diversity in archives? Drawing from interviews with academics, archivists, curators, and other experts across the UK/Europe and the USA, this article sheds light on the lack of collaboration between producers of AI technologies on the one side, and archivists, librarians and other cultural heritage professionals on the other side. We argue that bringing these stakeholders together is essential to improve the diversity of archival collections, using ethical and responsible AI. Finally, we offer recommendations to help professionals in libraries and archives assess the opportunities and risks associated with AI and find solutions to make their collections more representative of diverse audiences.</p></div>","PeriodicalId":47165,"journal":{"name":"AI & Society","volume":"40 6","pages":"4447 - 4459"},"PeriodicalIF":4.7,"publicationDate":"2025-02-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s00146-025-02222-z.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144909677","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}