Newsnet #5. The impact of artificial intelligence on social communication

Newsnet #5 The Impact of Artificial Intelligence on Social Communication December 13th, 2024 http://go.ehu.eus/newsnet Newsnet #5 (Bilbao, December 13th, 2024) THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION NEWSNET #5 SEMINAR REPORT Simón Peña-Fernández Koldobika Meso-Ayerdi Ainara Larrondo-Ureta (Eds.) CIP. Biblioteca Universitaria Newsnet Seminar (5º. 2024. Bilbao) The impact of artificial intelligence on social communication [Recurso electrónico]: Newsnet #5 Seminar Report / Simón Peña-Fernández, Koldobika MesoAyerdi, Ainara Larrondo-Ureta (Eds.). – Datos. – [Leioa] : Universidad del País Vasco / Euskal Herriko Unibertsitatea, Argitalpen Zerbitzua = Servicio Editorial, [2025]. – 1 recurso en línea : PDF (108 p.) Modo de acceso: World Wide Web. Precede al tít.: Newsnet #5 (Bilbao, December 13th, 2024) ISBN: 978-84-1319-690-9 1. Periodismo en línea. 2. Medios de comunicación social. 3. Inteligencia artificial. 4. Desinformación. I. Peña Fernández, Simón, ed. II. Meso Ayerdi, Koldo, ed. III. Larrondo Ureta, Ainara, ed. (0.034)316.77: 004.8 (0.034) 070:004.8 Newsnet “Impact of artificial intelligence and algorithms on online media, journalists and audiences” (PID2022-138391OB-I00) and “Automated counter narratives against misinformation and hate speech for journalists and social media” (TED2021-130810BC22) and are research projects funded by the Spanish Ministry of Science, Innovation and Universities and by the European Commission NextGeneration EU/PRTR. Irati Agirreazkuenaga-Onaindia, Maider Eizmendi-Iraola, María Ganzabal-Learreta, Ainara Larrondo-Ureta, Terese Mendiguren-Galdospin, Koldobika Meso-Ayerdi, Julen Orbegozo-Terradillos, Simón Peña-Fernández, Carmen Peñafiel-Saiz, Jesús Ángel Pérez-Dasilva, Reyes Prados-Rodríguez Gureiker Research Group of the Basque University System (A) (IT 1496-22) © Servicio Editorial de la Universidad del País Vasco Euskal Herriko Unibertsitateko Argitalpen Zerbitzua ISBN: 978-84-1319-690-9 NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION INDEX Simón Peña-Fernández, Koldobika Meso-Ayerdi & Ainara Larrondo-Ureta Preface ARTIFICIAL INTELLIGENCE IN JOURNALISM AND COMMUNICATION Mathias-Felipe de-Lima-Santos AI to Empower Media and Democracy in the Global South: A Critical Reflection Javier Díaz-Noci Transparency when Using Artificial Intelligence: A Copyright Law Approach Ainara Larrondo-Ureta, Simón Peña-Fernández & Mikel Leibar AI, Media and Journalism: Editorial and Algorithmic Decisions in Public Service Media (PSM) Terese Mendiguren-Galdospin, Koldobika Meso-Ayerdi, Reyes PradosRodríguez & Urko Peña-Alonso Artificial Intelligence Applied to Informative Speech: An Exploratory Study with Journalism Students Javier Odriozola-Chéne, Rosa Pérez-Arozamena & Javier Díaz-Noci Artificial Intelligence, Journalistic Production, and Media Consumption: The Design of a Methodological Tool for the Analysis of Related Media Coverage in Spain Barbara Sarrionandia Disinformation and Gender Bias in Artificial Intelligence: Analysis, Impact, and Proposals for Technological Equity NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION Matthew Quick Coming of Age: Drivers and Consequences of Artificial Intelligence on the News Publishing Economy Elena Yeste-Piquer & Jaume Suau-Martínez Citizen Perception of Artificial Intelligence: Impact and Applications in Journalism RESEARCH ON DIGITAL COMMUNICATION Maider Eizmendi-Iraola, Ainara Larrondo-Ureta, Ainhoa Novo-Arbona, Julen Orbegozo-Terradillos & Simón Peña-Fernández Basque Feminism on Social Media Kyle Leaver Media Desensitization in Coverage of Mass and School Shootings: State of the Art and Future Research Proposal Jesús Ángel Pérez-Dasilva, Maria Ganzabal-Learreta & Urko Peña-Alonso Breast Cancer on TikTok: Emotional Support, Misinformation, and the Challenge of Reliable Health Communication Mayte Santos-Albardia Critical Media Literacy and Migration in Higher Education: A DialogicalCritical Approach NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION PREFACE The expansion of generative artificial intelligence has transformed the way information is produced and circulated, raising pressing questions about its impact on truthfulness, authorship, and informational responsibility. In this new landscape, disinformation is taking on increasingly sophisticated and harder-todetect forms, making its study an urgent task in understanding the balance between technology, power, and public knowledge. In this context, the Newsnet #5 seminar took place at the University of the Basque Country (UPV/EHU) on December 13, 2024 and sought to find answers to understand the impact that the incorporation of artificial intelligence will have in the media, particularly in aspects such as misinformation or the impact on journalists and their audiences. This event is organised as part of the activities carried out within the framework of the projects Impact of artificial intelligence and algorithms on online media, journalists and audiences (PID2022-138391OB-I00) and Automated counter narratives against misinformation and hate speech for journalists and social media (TED2021130810B-C22). These projects, funded by the Spanish Ministry of Science, Innovation and Universities, aim to analyse how misinformation spreads and how the media seek to counter it, as well as to assess the impact of artificial intelligence implementation, particularly in its social dimension. This volume is structured in two main sections that are in dialogue with one another. The first examines how artificial intelligence is becoming embedded at the core of communication and journalism, reshaping the ways in which information is produced, distributed and received. Through various studies, it explores processes of automation, and the tensions that arise between technological innovation and the principles underpinning journalistic practice. It also addresses the ethical, legal and social challenges that accompany this transition. NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION The second section broadens the focus to other forms of digital communication. It brings together research that investigates how social media and new expressive formats are redefining the relationships between media and their audiences. These contributions allow for the situating of artificial intelligence within a wider ecosystem, where technology, culture and everyday life intersect. Taken together, the chapters in this volume provide a broad and detailed picture of the challenges and possibilities that artificial intelligence brings to the field of social communication. They aim to encourage critical thinking about how these technologies are shaping public debate and the role of journalism in today’s digital environment. The editors wish to express their gratitude to the authors of the contributions, as well as to all the individuals and institutions whose support has made these projects possible. Simón Peña-Fernández Koldobika Meso-Ayerdi Ainara Larrondo-Ureta Editors NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 8 ARTIFICIAL INTELLIGENCE IN JOURNALISM AND COMMUNICATION NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 9 AI TO E MPOWER M EDIA AND D EMOCRACY IN THE G LOBAL SOUTH: A CRITICAL REFLECTION Mathias-Felipe de-Lima-Santos [email protected] Pompeu Fabra University (UPF) I NTRODUCTION Artificial Intelligence (AI) is reshaping the global media landscape (de-LimaSantos & Ceron, 2021). In the Global South— regions often characterized by economic limitations, political instability, and infrastructural challenges (Dados & Connell, 2012)—this transformation presents both unprecedented opportunities and serious risks. This essay, based on the presentation “AI to Empower Media and Democracy in the Global South” at the University of the Basque Country (UPV/EHU), provides an overview of AI use in Global South news outlets. It highligh ts how these organizations navigate pressing challenges like digital inequality, media freedom constraints, institutional isomorphism, and the tension between innovation and responsible adoption. This essay critically reflects on these key themes, contextualizing them within the Global South’s broader socio-political realities and the scholarly discourse on AI in news media. THE GLOBAL SOUTH CONTEXT: A COMPLEX LANDSCAPE This discussion is grounded in the complex realities of the Global South, comprising countries historically facing systemic marginalization in global economic and technological developments. Media systems in these regions frequently operate amid limited press freedom, widespread corruption, and insufficient access to technological resources (Freije, 2021). These conditions NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 16 for journalists, researchers, and citizens interested in governmental transparency, leveraging technology to make public data more accessible and understandable. - GMA Integrated News Digital in the Philippines released “InoculatED,” which uses AI to proactively combat misinformation through “prebunking”— educating audiences about misinformation tactics before they are exposed to them. Central to this initiative is Facts Talk, a series of engaging video explainers produced with the help of generative AI tools. The integratio n of AI has significantly improved efficiency, cutting production time by 50% and doubling the output of vertical videos for platforms such as YouTube Shorts. These cases demonstrate that, when localized and ethically deployed, AI can significantly enhan ce media’s capacity to serve the public interest. They highlight the importance of integrating contextual knowledge and community engagement in technological design—something often overlooked in Global North-centric AI systems. CONCLUSION AI has the pote ntial to significantly empower media systems in the Global South— enhancing transparency, increasing efficiency, and expanding democratic participation. However, this potential can only be realized if the deployment of AI is context-sensitive, ethically inf ormed, and inclusive. Structural barriers such as the digital divide, institutional isomorphism, and global power asymmetries must be addressed to ensure that AI becomes a tool for empowerment rather than exclusion. Collaboration between journalists, tech nologists, civil society, and academic institutions is key to developing AI systems that reflect local languages, cultures, and political dynamics. Additionally, continuous evaluation of AI’s NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 17 impact on newsroom practices and democratic outcomes is essential to ensure that innovation does not undermine journalistic integrity. By critically assessing both the risks and opportunities of AI integration, media practitioners and policymakers in the Global South can build more resilient, innovative, and democratic information ecosystems. REFERENCES Cheas, K. (2024). Psychological capital and safety in Global North-South cooperation: A field analysis of collaborative investigative journalism across the U.S.-Mexico border. Journalism, 26(4), 781–799. https://doi.org/10.1177/14648849241286857 Coeckelbergh, M. (2024). Why AI undermines democracy and what to do about it. John Wiley & Sons. Dados, N., & Connell, R. (2012). The Global South. Contexts, 11(1), 12–13. https://doi.org/10.1177/1536504212436479 DalBen, S., & Jurno, A. (2021). More than code: The complex network that involves journalism production in five Brazilian robot initiatives. ISOJ Journal, 11(1), 111–137. https://isoj.org/research/more-than-code-thecomplex-network-that-involves-journalism-production-in-fivebrazilian-robot-initiatives/ de-Lima-Santos, M.-F., & Ceron, W. (2022). Artificial Intelligence in News Media: Current Perceptions and Future Outlook. Journalism and Media, 3(1), 13–26. https://doi.org/10.3390/journalmedia3010002 de-Lima-Santos, M., Munoriyarwa, A., Elega, A., & Papaevangelou, C. (2023). Google News Initiative’s Influence on Technological Media Innovation in Africa and the Middle East. Media and Communication, 11(2), 330– 343. https://doi.org/10.17645/mac.v11i2.6400 NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 18 de-Lima-Santos, M. F., Yeung, W. N., & Dodds, T. (2024). Guiding the way: A comprehensive examination of AI guidelines in global media. AI & Society, 1–19. https://doi.org/10.1007/s00146-024-01973-5 de-Lima-Santos, M.-F., & Jamil, S. (2024). Bridging the AI Divide: Human and Responsible AI in News and Media Industries. Emerging Media, 2(3), 335–346. https://doi.org/10.1177/27523543241291229 Freije, V. (2021). Selling democracy and press freedom to the Third World. Diplomatic History, 45(1), 72–82. https://doi.org/10.1093/dh/dhaa081 Gondwe, G. (2023). CHATGPT and the Global South: how are journalists in sub-Saharan Africa engaging with generative AI? Online Media and Global Communication, 2(2), 228–249. https://doi.org/10.1515/omgc-20230023 Jamil, S. (2022). Evolving Newsrooms and the Second Level of Digital Divide: Implications for Journalistic Practice in Pakistan. Journalism Practice, 17(9), 1864–1881. https://doi.org/10.1080/17512786.2022.2026244 Lelo, T. (2022). The Rise of the Brazilian Fact-checking Movement: Between Economic Sustainability and Editorial Independence. Journalism Studies, 23(9), 1077–1095. https://doi.org/10.1080/1461670X.2022.2069588 Mabweazara, H. M. (2020). Towards reimagining the ‘digital divide’: impediments and circumnavigation practices in the appropriation of the mobile phone by African journalists. Information, Communication & Society, 24(3), 344–364. https://doi.org/10.1080/1369118X.2020.1834602 Mesquita, L., & de-Lima-Santos, M. F. (2024). Google news initiative innovation challenge in Latin America: business models between path dependence and power relations. Journal of Media Business Studies, 1– 26. https://doi.org/10.1080/16522354.2024.2402630 Munoriyarwa, A., Chiumbu, S., & Motsaathebe, G. (2021). Artificial Intelligence Practices in Everyday News Production: The Case of South NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 19 Africa’s Mainstream Newsrooms. Journalism Practice, 17(7), 1374–1392. https://doi.org/10.1080/17512786.2021.1984976 Nishal, S., & Diakopoulos, N. (2023). Envisioning the applications and implications of generative AI for news media. In CHI ’23 Workshop on Generative AI and HCI (pp. 1–8). ACM. https://arxiv.org/pdf/2402.18835 Paulussen, S. (2016). Innovation in the newsroom. In T. Witschge, C. W. Anderson, D. Domingo, & A. Hermida (Eds.), The SAGE handbook of digital journalism (1st ed., pp. 192–206). SAGE. https://doi.org/10.4135/9781473957909.n13 Starke, C., Baleis, J., Keller, B., & Marcinkowski, F. (2022). Fairness perceptions of algorithmic decision-making: A systematic review of the empirical literature. Big Data & Society, 9(2). https://doi.org/10.1177/20539517221115189 Stilgoe, J., Owen, R., & Macnaghten, P. (2013). Developing a framework for responsible innovation. Research Policy, 42(9), 1568–1580. https://doi.org/10.1016/j.respol.2013.05.008 NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 20 T RANSPARENCY WHEN U SING A RTIFICIAL I NTELLIGENCE : A COPYRIGHT LAW APPROACH Javier Díaz-Noci [email protected] Pompeu Fabra University (UPF) I NTRODUCTION Media newsrooms have started to cover the popularization of artificial intelligence systems across the internet, which began with the public availability of ChatGPT, the flagship product of the Ame rican company OpenAI in November 2022, followed in February 2025 by the Chinese product DeepSeek and the European (French) Le Chat, from the company Mistral. The unrelenting adoption of these tools has led to at least two problems related to intellectual property and copyright law in the media industry and journalism. One is related to how training artificial intelligence models such as those mentioned above (and others), based on large language models (LLMs), requires huge amounts of data. This has raised serious questions, ending in court in more than a few cases, regarding whether such AI systems are using thirdparty works protected by intellectual property law, in many cases without citing the source. The second problem concerns the products of these AI systems. The issue of transparency is related to, and perhaps even lies at the heart of, both of these problems. This concept was initially applied, for example, in the EU Media Freedom Act, through the requirement that the media be transparent about the ownership of their companies as well as their sources of funding, including subsidies, institutional advertising they may receive, and loans from NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 21 financial institutions that may condition their editorial line and independence (Tomaz, 2024; Theina & Sevignani, 2024), but the same term also applies to attribution rights as established by intellectual property law. We refer herein to this latter meaning. CURRENT STATE OF AFFAIRS The influence of AI systems has led to legislation to specifically regulate AI, in systems that are normally based on risk such as the EU AI Act—an approach that is also followed in bills in Brazil (draft Lei da Inteligência Artificial, PL 2.338/2023, introduced in the Senate in December 2024) and Australia (Safe and responsible AI in Australia: Proposals paper for introducing mandatory guardrails for AI in high-risk settings, September 2024)—in addition to court cases and agreements between media and AI companies in parallel. However, concerns remain about how to ensure that transparency is maintained in all processes involving generative artificial intelligence. Transparency regarding the algorithm and its legal regulation is considered to be a central issue, being not only a legal but also an ethical requirement (Alén-Savikko, 2022). Shortly after the appearance of ChatGPT in November 2023, the G7 countries (USA, Canada, UK, France, Germany, Italy, and Japan) launched the Hiroshima Process International Guiding Principles for Organizations Developing Advanced AI Systems, leading to a code of conduct for AI developers based on voluntary adherence. These principles are aimed at identifying risks, insisting on the need for transparency and accountability, and ensuring both the privacy of personal data and intellectual property rights. Just one month later, 29 countries, including many from the European Union as well as the USA and China, adopted the Bletchley Declaration at a global summit on AI. Similarly, on October 30, 2023, the US Government issued an Executive Order on the Safe, Secure, and Reliable Development and Use of Artificial Intelligence. NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 22 Also in 2023, Digital Content Next, whose partners include some of the world’s most powerful newspaper publishers, released the Principles for Development and Governance of Generative AI, which insists on both intellectual property (of media publishers, rather than journalists) and, in point four, transparency as an obligation to clarify how generative artificial intelligence models have been trained. The development of such a statement of principles has also been adopted by the International Authors’ Forum (Principles for Artificial Intelligence and Authorship, September 2023), the first being authorization, fair compensation, and transparency (as well as accountability, see Milosavljević & Poler, 2024). This requirement for transparency has become law, although it will not really come into force until mid-2026, through the European AI Act. Thus, under EU law, article 104 states that one of the transparency requirements for AI system companies is to clarify the content that they use to train their models. In the final stage before the approval of this Act, journalists’ organizations such as Reporters Without Borders (RWB) recommended the inclusion of specific legal provisions. One provision that was proposed is that “content generated by large language models even in the supervised training phase must be verified by media and information professionals.” Another aimed to ensure that any source used for this purpose should uphold the principles of pluralism. Moreover, RWB urged "the European Commission to ensure that the General Purpose Artificial Intelligence (AI) Code of Practice includes specific provisions to protect journalism and reliable information," considering that 78% of the meetings held by European representatives were with people from the tech industry. This dominant position of the tech industry and the resulting danger to pluralism have also been denounced by the European Broadcasting Union (EBU, 2024, Borchardt et al., 2024) on behalf of the majority of European public broadcasting services. Specifically, RWB insisted on “improving transparency in the use of copyrighted works to train AI models, and ensuring more effective NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 23 means for authors and rights holders—including journalists and media organizations— to negotiate and retain their rights.” None of these recommendations were accepted or finally adopted in either the IA Act or the Code of Practice. Moreover, recent academic literature, while recognizing that transparency is an important value for journalists and media as well as for the public (Thomson et al., 2025: 9), states that “transparency is not a panacea,” and that, for example, Art. 50 of the EU IA Act is too narrow to be of any real benefit to address the needs of audiences (Piasecki, Morosoli, Helberger & Naudst, 2024). The News/Media Alliance (NMA) reiterated these principles in April 2023, expanding on them in their response to and comments during the consultation on Artificial Intelligence and Copyright launched by the US Copyright Office in October of that year. Finally, also in November 2023, the Paris Charter on AI and Journalism was launched. In contrast to the above-mentioned guidelines, it insists on ethical rather than legal solutions, but still calls for transparency (in particular to distinguish human from synthetically produced content) and accountability as the pillars of a human-centered perspective on the use of artificial intelligence in newsrooms. Other regulators have echoed this need for transparency regarding the use of sources. Even the House Bipartisan Task Force on Artificial Intelligence of the US Congress has been working in this issue, acknowledging in a report published in December 2024 that "it is often difficult for creators to know whether their copyrighted works are being used by AI developers.” Baroness Kidron’s proposed amendments to Clause 132 of the UK Data (Use and Access) Bill in the House of Lords seek to enforce precisely this transparency by demanding that all works and authors used in LLM training processes be explicitly acknowledged. The UK government also uses this argument, albeit from a different point of view; indeed, its December 2024 open consultation document on Copyright and Artificial Intelligence states that the proposed modifications (in fact, a licensing exemption) “depend on greater trust between NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 24 AI developers and rights holders.” However, it is unclear whether this proposed provision will ultimately be enacted. CONCLUSIONS AND FUTURE RESEARCH Space limitations prevent us from explaining herein the full importance and development of the concept of transparency in copyright law. For the time being, although more in the realm of wishful thinking than true regulatory effectiveness, transparency of sources and processes remains a central concept. The most advanced legislative initiative, apart from the already approved EU AI Act, is currently the Brazilian artificial intelligence law. This bill devotes an entire section to transparency and copyright, and if passed, all AI system developers will have to report the use of content protected by copyright and related rights and refrain from using content to which an opt-out has been applied. The Brazilian legislator, benefiting from others’ experiences, recognizes that it will not be easy to protect these rights in each and every work, and moreover, such control will be regulated by implementing legislation (regulations) and entrusted to an administrative authority (Rocha de Sousa, 2024: 94–96). The Australian government is proceeding in a similar manner, devoting the entire section 6 of its Proposals Paper for Introducing Mandatory Guardrails for AI in High-Risk Settings of 2024 to this topic. Australia is also a unique case because it has neither a fair-use exception nor a system of graded exceptions as in the EU. The practical and legal embodiment of the principle of transparency, especially regarding sources and the right to author attribution (which is not always given to journalists in all legal traditions, in particular being omissible but generally followed in Common Law countries such as the UK or Ireland) will be an issue to follow and analyze in the near future, considering the wide range of solutions and its implications for other branches of law, relating for example to competition, to privacy, or even to plagiarism. A certain laxity in this regard has led, for example, to the considerable proliferation in China of media NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 25 content reproduced without authorization or citation of origin or authorship via social networks (Xu, YU & Chen, 2024). This problem is not unique to China. In France, 40 newspapers went to court in February 2025 to prevent systematic plagiarism, using artificial intelligence tools, of their content by web portals such as NewsDayFR. The emergence of DeepSeek in February 2025 provoked reactions and even measures from authorities such as those in Italy or Ireland, precisely because of misgivings regarding the processing of personal data, which may be in contradiction to European legislation, as well as other concern s closer to journalistic practice, such as the possibility of copyright infringement (Adami, 2025; Deck, 2025). One of the solutions comes through selfregulation by the media and newsrooms, and it is possible that this trend will become more pronounced in the near future. Media such as El País in Spain or the British (actually, global) The Guardian publicly stated in 2024 and 2025 that they will have such internal regulation in place. It is taken for granted that measures such as this will reinforce users’ trust in the media (Linden & Tuulonen, 2019; Mitchell, 2025). At the same time, The Guardian, one of the last to sign an agreement with OpenAI, in February 2025, states it can thereby guarantee attribution of content used by ChatGPT. No actions, even the most extreme, can be ignored, such as the appeal to the courts by the French Syndicat d’Editeurs de Presse Magazine, which succeeded in November 2024 in getting the Paris Commercial Court to prohibit Google from removing news articles that came from certain media with which it had copyright disagreements. The Authors Guild has published a seal of quality, that is, th e Human Authored Certification, that guarantees that content reaching a user was created through decisive human intervention. To the best of our knowledge, at the time of writing, no media had joined this initiative. In contrast, we have seen the appearance of media outlets such as Quartz that, without acknowledging it, have published numerous AI-generated news stories. Tools such as SpinozIA, from Reporters Without Borders, aim to provide journalists NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 32 characterized PSMs, which seem difficult to express via a mathematical formula (Sørensen, J. K.; Hutchinson, 2018). In this vein, the values of PSMs are already facing technical and organizational drivers that may pro mote their profound transformation, starting with a transfer of autonomy from editorial to technical staff (Carillon, 2024). So, the introduction of AI causes tensions between editorial considerations and algorithmic decisions in an organizational and labo r context in which journalists constitute only a fraction of the constellation of actors involved in the development of recommendation systems. Indeed, engineers, designers, or decision-makers exert a much more significant influence on these processes. These circumstances also make it necessary to consider the type of intelligent systems implemented at each broadcaster, i.e., whether to opt for the development of proprietary algorithms, with the associated economic costs, or to use third-party products. In this case, in particular, it becomes essential to monitor the type of data that feeds the algorithms, to avoid biases or inappropriate data use and ultimately ensure the quality of the product resulting from the use of such data, in terms of its usefulness to audiences. Indeed, some authors state that the use of third-party applications or large technological platforms is not ideal, “especially in the case of public service media, owing to their specific content and data processing needs” (FieirasCeide, Vaz-Álvarez and Túñez-López, 2022). CONCLUSIONS This opens up an interesting line of research into the strategies employed by broadcasters to integrate public service values into content recommendation systems. This line of work has begun to be explored through the work of the Gureiker Research Group (IT149622) in its collaboration with the Aula Empresa of the Basque Public Radio and Television EITB and the Faculty of Social Sciences and Communication of the UPV/EHU, the university to which NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 33 the group is attached. On the basis of the analysis of experiments (Orbegozo et al., 2025), cases (Leibar, 2024), and following the work of technical staff, the importance of paying more attention than ever to technical issues is emphasized, but with the aim of discovering the current potential regarding news and journalism. This line of research also involves the transfer of university teaching of relevant journalism competences to train future communicators in the design, programming, and use of algorithms while incorporating the ethical and professional values of journalism. REFERENCES Aramburú Moncada, L. G., López Redondo, I., & López Hidalgo, A. (2023). Inteligencia artificial en RTVE al servicio de la España vacía. Proyecto de cobertura informativa con redacción automatizada para las elecciones municipales de 2023. Revista Latina de Comunicación Social, 81, 1-16. https://www.doi.org/10.4185/RLCS-2023-1550 Carillon, K. (2024). “An Algorithm for Public Service Media?” Embedding Public Service Values in the News Recommender System on RTBF’s Platform. Emerging Media, 2(3), 422-448. https://doi.org/10.1177/27523543241290976 Díaz-Noci, J. (2023). Legal problems (and some remedies) related to artificial intelligence, media and copyright law: a state of the art (so far). Newsnet#4: The Challenges of Artificial Intelligence for Journalism. Leioa: University of the Basque Country. European Audiovisual Observatory (2024). AI and the audiovisual sector: navigating the current legal landscape. Estrasbourg. https://rm.coe.int/iris-2024-3-ia-legal-landscape/1680b1e999 Fieiras-Ceide, C., Vaz-Álvarez, M., & Túñez-López, M. (2022). Artificial intelligence strategies in European public broadcasters: Uses, forecasts NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 34 and future challenges. Profesional de la Información, 31(5). https://doi.org/10.3145/epi.2022.sep.18 Fieiras-Ceide, C., Vaz-Álvarez, M.; Túñez-López, M. (2024). Designing personalisation of European public service media (PSM) trends on algorithms and artificial intelligence for content distribution. Profesional de la información, 32(3), https://doi.org/10.3145/epi.2023.may.11 García-Avilés, J. A. (2021). Journalism innovation research, a diverse and flourishing field (2000—2020). Profesional de la información, 30(1), 1-25. https://doi.org/10.3145/epi.2021.ene.10 Leibar, M. (2024). Aplicación de la Inteligencia Artificial (IA) para la creación de contenidos audiovisuales: análisis del caso “Hiperia” de RTVE [Master’s Dissertation]. UPV/EHU. Peñafiel-Saiz, C., Peña-Fernández, S., & Larrondo-Ureta, A. (2024). Oportunidades y desafíos de la Inteligencia Artificial en el marco legal europeo, 15-37. In Murcia-Verdú, J., & Ramós-Antón, J. (2024). La Inteligencia Artificial y la transformación del periodismo Narrativas, aplicaciones y herramientas. Comunicación Social. Public Media Alliance (2023). How Public Media is adopting AI https://www.publicmediaalliance.org/how-public-media-is-adoptingai/ Sørensen, J. K., & Hutchinson, J. (2018). Algorithms and Public Service Media, 91-106. In Public Service Media in the Networked Society: RIPE@2017. Nordicom. http://www.nordicom.gu.se/sites/default/files/publikationerhelapdf/public_service_media_in_the_networked_society_ripe_2017.p df NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 35 A RTIFICIAL I NTELLIGENCE A PPLIED TO I NFORMATIVE S PEECH : AN EXPLORATORY STUDY WITH JOURNALISM STUDENTS Terese Mendiguren-Galdospin terese.mendigu[email protected] University of the Basque Country (UPV/EHU) Koldobika Meso-Ayerdi [email protected] University of the Basque Country (UPV/EHU) Reyes Prados-Rodríguez [email protected] University of the Basque Country (UPV/EHU) Urko Peña-Alonso [email protected] University of the Basque Country (UPV/EHU) I NTRODUCTION In the emerging context of the use of voices generated by artificial intelligence (AI) in the field of journalism, this study proposes a comparison and reflection on the differences between these and human voices. In particular, we address aspects such as credibility and emotional connection, which are key elements in media communication. In addition, we reflect on the extent to which AIgenerated speech could be considered to provide a valid substitute for human voices in the near future. This reflection is based on an experiment designed to evaluate the ability of journalism students at the University of the Basque NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 36 Country to differentiate between human and AI-generated voices. This exploratory study was carried out in the context of a classroom activity involving 100 journalism students from three different groups, considering specific voice factors that influence their differentiati on, such as intonation, rhythm, timbre, and vocalization. JOURNALISM, SPEECH, AND ARTIFICAL INTELLIGENCE The introduction of AI voice generation raises profound questions and challenges regarding the future of communication. Technological innovations make room for “diverse typologies, such as audio fiction, narrative non-fiction, conversational, or popularization podcasts” [“tipologías diversas, como ficciones sonoras, podcasts narrativos de no ficción, conversacionales o de divulgación”] (Chaparro Doming uez, 2024: 120) and a range of other automation resources, allowing the media to experiment extensively (BazanGil et al, 2021). Among the examples that illustrate artificial innovations to conquer the audio field is the case of Intar Radio, the first online radio station to be created in Spain with AI. Although this project is novel, it suffers from monotonous prosodic features that limit its emotional connection with the audience (Gómez, 2024). Among the mainstream media, Cadena SER incorporated the voice of an artificial announcer (Victoria) into its program Carrusel Deportivo in 2022, an experimental multiplatform initiative that has also been applied by Amazon with Alexa and in the As daily newspaper (Novoa, 2023). The voice is an essential component of the human experience; it transmits not only information but also emotions, intentions, and nuances that enrich the message (Rodero, 2003), making it one of the most complex characteristics of the human body (Torres Gallardo, 2025). In this regard, a natural voice still has much greater capacity than an artificial voice to generate empathy and establish an emotional bond with the audience (Gómez, 2024), aspects that are fundamental in journalism (Fitó Carreras et al., 2023). Meanwhile, AI- NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 37 generated voices, although they have advanced considerably in terms of technical quality, “have not yet managed to fully mimic the many nuances of the human voice” [“todavía no han conseguido imitar completamente los numerosos matices que tiene la voz humana”] (Chaparro-Dominguez, 2024: 135) and thus face challenges in achieving the same level of connection and authenticity. In this context, and considering that AI development is advancing rapidly, the similarity of artificial to human voices will probably soon improve (Jimenez et al, 2024). It is thus considered interesting to test the human perception of the authenticity of different voices, both natural and artificial, in an audio news context, where the emotional nuances related to spontaneity are more limited than in entertainment contexts. METHODS This study aims to explore how AIgenerated voices are perceived in comparison with human voices, primarily from a prosodic perspective. For this purpose, an experiment was designed in which 96 students following different journalism courses listened to 14 audio fragments with the same journalistic content: 7 produced by human speakers and 7 by free AI applications available on the market. The participants had to identify which voices were natural and which were artificial. The experimental design included a semi-open-ended questionnaire in which specific elements that influence voice perception were to be evaluated. These factors included: intonation, which refers to tonal variations in speech; rhythm, related to the spee d and fluidity of speech; timbre, which determines the characteristic “color” of each voice that gives it its personality; and vocalization, which describes the precise and articulate pronunciation of sounds. These elements are fundamental to transmit not only the semantic content of the message but also its intention and emotionality. In addition, students responded to questions related to the implementation of AIgenerated speech in journalism. NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 38 R ESULTS Faced with the challenge of having to determine whether the played-back audios were produced by humans or generated by AI, the respondents did not achieve unanimous consensus for any of the 14 audios. It it also interesting to note that the greatest margin of doubt was generated by the artificial voices. Figure 1. Perception of veracity of AI-generated audio Source: Authors’ own creation Although most of the artificial speech samples were recognized as such with almost 100% accuracy, one of them was detected by only 54.2% of the participants (Figure 1). This means that almost half believed it to be a natural voice. On the other hand, the human voice that generated the most doubt was recognized as natural by 77.1% of the respondents (image 2). In all other cases the margin was very small, indicating that the difference was easily detected. Meanwhile, the idea that all the voices managed to mislead their listeners in the some proportion is noteworthy, since none of them received a 100% unanimous decision. For all but two of the audios, the prosodic characteristic that most stood out as a determining factor in their perception of vocal realism was intonation. In one case, corresponding to an AIgenerated audio, the rhythm was highlighted, whereas in another case, being a natural voice, the characteristic that enabled its detected as natural was the vocalization. In any NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 39 case, all the voice recordings managed to “fool” the participants in this study to some extent, albeit limited. Figure 2. Perception of veracity of AI-generated audio Source: Authors’ own creation All respondents expressed an a priori preference for the human voice over AIgenerated voices. Among the main reasons expressed by the participants, the one related to expressiveness stands out. The human voice is considered to be able to convey emotions more effectively than an artificial voice. One of the recorded responses states that “if a person is moved by a topic, they will say it with more emotion; just as if it is a sad topic, they will say it with more regret, which the AI would have a harder time portraying.” [“si a una persona le emociona un tema, lo dirá con más emoción, igual que si es un tema triste, lo dirá con más pesar, cosa que a la IA le costaría más representar.”] The lack of naturalness and familiarity was also mentioned, sinc e the human voice “sounds more natural, inspires much more confidence, and is not boring, as the one generated with artificial intelligence seems very monotonous to me.” [“suena más natural, inspira mucha más confianza y no aburre, porque la generada con inteligencia artificial me parece muy monótona.”] Other participants noted that the human voice “is more familiar, as it allows you to recognize the voice of the person who is speaking” [“es más cercana, ya que te permite reconocer la voz de la persona que está hablando”]. With regard to the professional or employment dimension, statements such as “I wouldn't want NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 40 to listen to an artificial intelligence deliver the news” [“no me gustaría escuchar a una inteligencia artificial dar las noticias”] or “many people could be left on the street” [“mucha gente se podría quedar en la calle”] were mentioned. Another respondent argued: “We cannot lose the essence of communicating among human beings without relying on external agents” [“No podemos perder la esencia de comunicarnos entre los seres humanos sin depender de agentes externos”]. Regarding the limitations of AI, it was mentioned that, although AI-generated voices may improve over time, they currently still lack spontaneity. Meanwhile, 63.2% of respondents feared that, although artificial voices do not currently sound human, they will eventually do so and there is a possibility that human voices will be almost completely replaced. Indeed, more than half (56.1%) believed that, in the future, artificially generate d voices will not be detectable as such by simple listening. In addition, 73.7% of the students who volunteered for this exploratory study stated that they are somewhat or very concerned that, in the future, there will be less diversity of accents or dialects detectable in speech. CONCLUSIONS The results of this study reveal that current technologies for generating artificial voices still fail to fully replicate the authenticity and closeness of human voices. While modern technologies have significantly i mproved in aspects such as intonation and clarity, they often lack the emotional depth and dynamism of a human voice. AIgenerated voices tend to maintain a less natural intonation and rhythm than human voices or lack the subtle inflections that human voic es display to emphasize key points or express emotion. However, it is striking that all the voices recorded and played back in this study invoked a margin of doubt, even if minimal, regarding their authenticity. With technologies advancing at an unstoppabl e pace, this margin could increase significantly for speech generated by artificial intelligence in the future. NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 41 The inclusion of journalism students as participants in this experiment enabled us to explore their perceptions regarding the impact of AI on their future profession. More than half expressed concerns regarding how automation could reduce the presence of human speakers, and the resulting lack of vocal diversity that this trend may cause in the future. On the basis of this exploratory pre-analysis, future research will address the incorporation of more advanced deep learning models that more effectively reproduce not only linguistic patterns but also the prosodic and emotional signals present in human speech. On the other hand, it is important to consider the ethical and social implications of this technology. The use of AI-generated voices in journalism raises questions about transparency and authenticity. Should the media clearly inform their audience when using AI-generated voices? How might th is affect public confidence in the news? These are questions that require an open and ongoing debate among developers, journalists, and academics. REFERENCES Bazán-Gil, V., Pérez-Cernuda, C., Marroyo-Núñez, N., Sampedro-Canet, P., & De-Ignacio Ledesma, D. (2021). Inteligencia artificial aplicada a programas informativos de radio. Estudio de caso de segmentación automática de noticias en RNE. Profesional de la información, 30(3), e300320. https://doi.org/10.3145/epi.2021.may.20 Chaparro-Domínguez, M. Á. (2024). El impacto de la IA en los contenidos periodísticos sonoros. Espejo de Monografías de Comunicación Social, 25, 119-139. https://doi.org/10.52495/c5.emcs.25.p108 Fitó-Carreras, M., Vidal-Mestre, M., & Freire-Sánchez, A. (2025). Análisis de softwares de inteligencia artificial generativa de voz aplicados al podcasting. Comunicación y Hombre, 21, 179-196. NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 48 principles in the algorithm, resorting to the use and structure of topics derived from clickbait, disconnection from the interests of the audience), and speed (reduction of verification processes, an obstacle to deepening and contextualization, saturation of the audience due to information overload). Figure 2. Content analysis codebook BACKGROUND VARIABLES 1.ID Encoder 3. Coding date (dd/mm/yyyy) 4. URL 5.Date of publication (dd/mm/yyyy) * In case of update, enter the update date CONDITIONING FACTORS F OR THE JOURNALISM 6. Media 20minutos.es 1 elmundo.es 6 abc.es 2 elpais.com 7 elconfindencial.com 3 elperiodico.es 8 Eldiario.es 4 larazon.es 9 elespanol.com 5 okdiario.com 10 7. Media ideology Conservative 1 Liberal 2 Social-democrat 3 Far right 4 Authorship None 1 Agencies 2 In some media, although the heading of the text reads “Editorial team” [“Redacción”], at the end of the text it is indicated that it is an agency news item. Generic byline 3 For example: Editorial office, La Vanguardia... In-house journalists 4 Non-journalist collaborators 5 For example, guest columnists or opinion-formers. Could be an individual or collective author (e.g., Association of Spanish Geographers). NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 49 9. Journalistic function Informative 1 News, reports, interviews Interpretive 2 Reports, chronicles... Opinion 3 Columns, editorials, essays... 10. Main source Government agencies 1 Tech companies 2 Academics 3 Actors in the media sector (journalists, editors, etc.) 4 Non-governmental organizations and activists 5 Users 6 Other media 7 Other 8 None 9 DEFINITION OF AI IN THE CONTEXT OF THE MEDIA Type of AI 11. Generative AI: new content creation and content personalization No 0 Yes 1 12. Analytical AI: classification and analysis of data for predictions to support decision-making No 0 Yes 1 Relations hip of AI with the phases of the journalisti c productio n process 13. Use of AI in information retrieval: automation of information collection and documentation No 0 Yes 1 14. Use of AI in information production: automated content production No 0 Yes 1 15. Use of AI in information distribution: information distribution and relationship with the audience No 0 Yes 1 Relations hip of AI to the standard functions of journalis m 16. Effects on the role of informing the public about authorities and powerful actors No 0 Yes 1 17. Effects on the investigation and surveillance function No 0 Yes 1 18. Effects on the function of fostering social empathy No 0 Yes 1 NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 50 18. Effects on the role of providing a public forum for debate No 0 Yes 1 18. Effects on the role of serving as an advocate for diverse policies and viewpoints No 0 Yes 1 18. Effects on the function of defending democracy No 0 Yes 1 CHALLENGES IN THE USE OF AI ASSOCIATED WITH MEDIA ACTORS Challenge s for the industry 19. Economic viability of AI development No 0 Yes 1 20. Legal responsibility for generated content No 0 Yes 1 21. The creation of content about people No 0 Yes 1 22. Ethics in the use of data No 0 Yes 1 23. The transparency of algorithms No 0 Yes 1 Challenge s for profession als 24. The loss of the symbolic capital of journalists as mediators No 0 Yes 1 25. Job losses No 0 Yes 1 26. Adaptation of training for professionals No 0 Yes 1 27. The emergence of new professional roles No 0 Yes 1 28. New relationships with technical personnel No 0 Yes 1 Challenge s for the 29. Audience credibility of AI-generated content No 0 NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 51 audience Yes 1 30. Audience quality of AI-generated content No 0 Yes 1 31. Audience accessibility and amenability of AI-generated content No 0 Yes 1 BENEFITS AND DISADVANTAGES OF THE USE OF AI IN PRODUCTION PROCESSES Benefits 32. Accuracy: reduction of errors in the verification of information and in the drafting of information No 0 Yes 1 33. Accessibility: improved readability of journalistic pieces, adaptation to multiple platforms and to people with disabilities No 0 Yes 1 34. Diversity of information: detecting peripheral visions, improving the context through translation services, diversity of points of view (gender, political, linguistic) No 0 Yes 1 35. Relevance: issues important to social life, thought-provoking topics, improvement of personalization services No 0 Yes 1 36. Speed: acceleration of transcription and editing processes, improved distribution according to the routines of the audience No 0 Yes 1 37. Total benefits stated in the journalistic text Damage 38. Accuracy: reduction of data quality, proliferation of disinformation (fake news, deep fakes), suppression of the appropriate context for correct interpretation No 0 Yes 1 39. Accessibility: use of information from fake accounts, incorrect levels of personalization of information, widening of the digital divide, payment for access to information No 0 Yes 1 40. Diversity of information: biases and lack of transparency in algorithm configuration, excessive repetition of popular topics, proliferation of hate speech, development of echo chambers or information bubbles No 0 NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 52 Yes 1 41. Relevance: absence of journalistic principles in the algorithm, recurrence of use and structure of topics that lead to clickbait, disconnection from the audience No 0 Yes 1 42. Rapidity: reduction of verification processes, slowing down of indepth analysis and contextualization, saturation of the audience due to information overload No 0 Yes 1 43. Total benefits stated in the journalistic text Source: Authors’ own creation REFERENCES Gómez-Calderón, B., & Ceballos, Y. (2024). Periodismo e inteligencia artificial. El tratamiento de los chatbots en la prensa española. index.Comunicación, 14(1), 281–300. https://doi.org/10.62008/ixc/14/01Period Gil de Zúñiga, H., Goyanes, M., & Durotoye, T. (2024). A Scholarly Definition of Artificial Intelligence (AI): Advancing AI as a Conceptual Framework in Communication Research. Political Communication, 41(2), 317-334. https://doi.org/10.1080/10584609.2023.2290497 Deuze, M., & Beckett, C. (2022). Imagination, Algorithms and News: Developing AI Literacy for Journalism. Digital Journalism, 10(10), 19131918. https://doi.org/10.1080/21670811.2022.2119152 Diakopoulos N. (2019). Automating the news: How algorithms are rewriting the media. Harvard University Press. Ioscote, F., Gonçalves, A., & Quadros, C. (2024). Artificial Intelligence in Journalism: A Ten-Year Retrospective of Scientific Articles (2014–2023). Journalism and Media, 5, 873–891. https://doi.org/10.3390/journalmedia5030056 NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 53 Lin, B., & Lewis, S. C. (2022). The One Thing Journalistic AI Just Might Do for Democracy. Digital Journalism, 10(10), 1627–1649. https://doi.org/10.1080/21670811.2022.2084131 Nielsen, R. K. (2017). The One Thing Journalism Just Might do for Democracy: Counterfactual idealism, liberal optimism, democratic realism. Journalism Studies, 18(10), 1251–1262. https://doi.org/10.1080/1461670X.2017.1338152 Owsley, C.S., & Greenwood, K. (2024). Awareness and perception of artificial intelligence operationalized integration in news media industry and society. AI & Society, 39, 417–431. https://doi.org/10.1007/s00146-02201386-2 Parratt-Fernández, S., Mayoral-Sánchez, J., & Mera-Fernández, M. (2021). The application of artificial intelligence to journalism: an analysis of academic production. Profesional de la información, 30(3). https://doi.org/10.3145/epi.2021.may.17 Peña-Fernández, S., Meso-Ayerdi, K., Larrondo-Ureta, A., & Dí¬az-Noci, J. (2023). Without journalists, there is no journalism: the social dimension of generative artificial intelligence in the media. Profesional de la información, 32(2). https://doi.org/10.3145/epi.2023.mar.27 Porlezza, C. (2023). Promoting responsible AI: A European perspective on the governance of artificial intelligence in media and journalism. Communications, 48(3), 370-394. https://doi.org/10.1515/commun-20220091 Sánchez-García, P., Merayo-Álvarez, N., Calvo-Barbero, C., & Diez-Gracia, A. (2023). Spanish technological development of artificial intelligence applied to journalism: companies and tools for documentation, production and distribution of information. Profesional de la información, 32(2). https://doi.org/10.3145/epi.2023.mar.08 Scheffauer, R., Gil de Zúñiga, H., & Correa, T. (2024). Algorithmic News Versus Non-Algorithmic News: Towards a Principle based Artificial NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 54 Intelligence (AI) Theoretical Framework of News Media. Profesional de la información, 33(1). https://doi.org/10.3145/epi.2024.0009 Túñez-López, M., Toural-Bran, C., & Valdiviezo-Abad, C. (2019). Automatización, bots y algoritmos en la redacción de noticias. Impacto y calidad del periodismo artificial . Revista Latina de Comunicación Social, (74), 1411–1433. https://doi.org/10.4185/RLCS-2019-1391 NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 55 D ISINFORMATION AND G ENDER B IAS IN A RTIFICIAL INTELLIGENCE: ANALYSIS, IMPACT, AND PROPOSALS FOR TECHNOLOGICAL EQUITY Barbara Sarrionandia [email protected] University of the Basque Country (UPV/EHU) I NTRODUCTION Gender bias in AI is one of the most pressing issues in contemporary technological development. Research has highlighted the lack of diversity in both the development teams and the datasets used, resulting in systems that not only reproduce but also amplify pre-existing gender stereotypes (Leavy, 2018; Nadeem et al., 2020). This underscores the need to implement corrective measures such as the inclusion of gender perspectives and equity principles, starting from algorithmic design. On the other hand, digital platforms also act as amplifiers of misogynistic discourses and gender disinformation (Cabañes, 2020), negatively impacting women’s safety and well-being (Benjamin, 2019), which reinforces the urgency of analyzing how media literacy can empower audiences to confront these dynamics. As technology becomes a central element of daily life, concerns about its impact on gender equality are also growing (Liu, 2024). Algorithmic systems are not neutral but reflect the biases and limitations of those who design them and the data used to train them (Wellner, 2020; Buijsman and Jänicke, 2021). In this NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 56 context, it is crucial to examine the social effects of these technologies and propose strategies to mitigate them. METHODOLOGY AND RESEARCH HYPOTHESIS Three fundamental hypotheses are proposed herein: 1. AIgenerated disinformation perpetuates gender stereotypes in society. 2. Algorithmic biases negatively impact women’s career opportunities. 3. Gender media literacy contributes to reducing discrimination and disinformation, thus promoting equity. These hypotheses are analyzed through a systematic literature review based on the key concepts of “gender disinformation,” “gender bias,” and “artificial intelligence,” along with synonyms and related definitions (such as “gender disinformation,” “disinformation,” “gender bias,” “gender equality,” and “femtech”), both in the field of journalism and in the media. ANALISYS The gender biases that are present in artificial intelligence (AI) algorithms play a crucial role in reproducing discriminatory patterns inherent to the data used during their training. For example, cases of hiring systems that penalize women and facial recognition technologies that suffer from significantly higher error rates for women and racialized individuals have been documented (Crawford, 2013; Greenfield et al., 2018). These failures reflect not only technical deficiencies but also a lack of diversity in the teams responsible for the design and training of such systems. Cases such as Amazon's algorithm that automatically downgraded resumes from female applicants show how seemingly objective decisions can be imbued with social bias, thus reinforcing existing inequalities. These situations highlight the importance of implementing systematic audits, ensuring transparency in design processes, NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 57 and adopting ethical principles from the very early stages of technology development (Lambrecht and Tucker, 2016). On the other hand, the feminization of technologies reinforces gender stereotypes that perpetuate traditional roles. Virtual assistants such as Siri and Alexa, by incorporating female voices associated with service and care tasks, reflect a cultural construction that links the feminine with characteristics such as obedience and availability (Sutko, 2020; Costa & Ribas, 2019; UNESCO, 2019). Moreover, humanoid robots such as Sophia consolidate the objectification of women, thus h eightening technological and social inequalities (Zimmerman, 2018). These design decisions are not accidental but respond to cultural dynamics that reinforce subordinate behaviors in technologies, conditioning not only the interactions of users but also their perception of the role of women in technological fields (Nomura, 2020; Eyssel & Hegel, 2012). This phenomenon underscores how gender representations in technologies can impact both the design of tools and the narratives that legitimize structural inequalities. In this context, gender media literacy emerges as a key tool to counteract such dynamics. Promoting inclusive digital skills enables people to develop critical thinking that facilitates the identification of biases in technological systems, as well as ensuring and supporting women's and girls’ access to STEM education. According to UNESCO, education in this area is essential to move toward a more equitable society (García Matilla, 2015). This approach not only strengthens the ability of audiences to recognize biases but also drives their active involvement in questioning the structures that support them. In addition, ensuring that emerging technologies reflect equitable values can contribute to greater public confidence in their use, an essential aspect in our increasingly digitized world (García Matilla, 2015; López Safi, 2015; Buitrago, Navarro, & García Matilla, 2015). NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 64 Furthermore, Sjøvaag (2024) suggests that newsrooms adopt AI as a necessity rather than a choice, with large tech companies exerting technical and economic pressures. Sjøvaag states, “news media outsource more and more of their monetization to thirdparty services that link content with consumers through programma tic advertising. News media also largely outsource distribution to platforms relying on automated processes to reach audiences. Moreover, their audience data and content catalogs are housed by cloud platforms whose data optimization processes rely on artif icial intelligence.” (Sjøvaag, 2024, p. 252). Partnerships or Dependency? Direct partnerships between news publishers and AI developers are not so straight-forward. In 2023, German news publisher Axel Springer licensed their content to AI developer OpenAI to train their Large Language Model (LLM) on real-time copyrighted journalistic content in exchange for tens of millions of dollars (Soper, 2023). In that same year, the New York Times (NYT) sued OpenAI for copyright infringement claiming the developer trained their models on the newspaper’s content without permission (Robertson, 2024). However, the allure of using AI is too grand, even for the NYT, who has since directed its journalists to use OpenAI tools in their newsroom during litigation (Benton, 2025). Moreover, AIdriven news content lacks clear attribution, raising ethical concerns about authorship and accountability (Scheffauer et al., 2024) while also acting as a direct competitor in the sphere (Robertson, 2024). These concerns have led to further legal action, with the News/Media Alliance filing a copyright and trademark infringement case against Cohere Inc - an AI company valued at over $5 billion (News Media Alliance, 2025). AI is increasing news organizations' dependence on platform companies, while these companies have historically controlled news distribution, AI adoption NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 65 now extends their influence to news production itself, threatening journalistic autonomy (Simon, 2022). AI in the Newsroom Contemporary research indicates that journal ists and editorial staff are beginning to develop workflows involving GenAI (Wu, 2024). Journalists in Singapore are using GenAI to help develop headlines, generate questions for interviews, and identify trending keywords (Macalintal, 2024). Meanwhile, journalists in Dubai are using GenAI to “gather stories from raw data and turn these data into intelligent stories” (Ahmad et al., 2024). SUGGESTED METHODOLOGY There are a number of methodological approaches to research in the field of journalism. This study will use a mixedapproach of qualitative and quantitative methods including case study reviews and semi-structured interviews. Case Studies Case studies allow researchers to analyze specific implementations of AI in journalism, such as automated news rooms, to understand their practical impact (Carlson, 2015). Since AI in journalism is still evolving, case studies provide early insights into organizational strategies concerning AI implementations and economic shifts (de-Lima-Santos & Ceron, 2022). The planned research will examine three to five case studies focused on media organizations that have actively implemented AI or worked with platforms for AI solutions. Criteria selection includes: NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 66 − The extent of AI integration in content production, audience engagement, and monetization. − The economic models underpinning AI adoption, including partnerships with AI firms (e.g., Axel Springer’s deal with OpenAI). − The policies and editorial guidelines used to ensure AI-generated content maintains journalistic integrity. − Sources of data will include public reports, corporate documents, industry analyses, contemporary reporting, and prior research studies. Semi-Structured Interviews Data collection in this field occurs in a number of ways, most predominantly through semistructured interviews, as Cools and Diakopoulos (2024) state “semistructured interviews were conducted as a means of facilitating purposeful and focused conversations. These interviews offered the necessary flexibility to delve into the subjects’ insights, interests, and areas of expertise” (p.7). Interviews are a well established methodology that allows for rich qualitative data collection (Guion et al., 2011) which often requires gathering personal insights and experiences that could not be c aptured through quantitative methods (Albizu-Rivas et al., 2024). Furthermore, semi-structured interviews allow researchers to adapt questions based on the journalist’s responses, making it possible to probe further into areas of interest that emerged during conversations (Albizu-Rivas et al., 2024). Researchers have also used semistructured interviews to get a deeper understanding of how business leaders react to disruptions in the industry (Sjøvaag, H., & Owren, T. 2021), a topic this paper specifically aims to understand further. The study will conduct semistructured interviews with key stakeholders, including journalists, editors, and newsroom executives working at three NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 67 digital news media organizations within the United States of America (USA). Participants will be drawn from three newsrooms that differ in content and audiences they serve. The organizations will be as follows: − One organization serving news to a local audience such as a county newspaper − One organization serving news to a national au dience covering the United States − One organization serving news to a global audience, similar to The New York Times, Business Insider, or Forbes The interview questions will focus on: − Perceived economic and editorial benefits and challenges of AI in the newsroom − Strategies to adopt AI to increase revenue or introduce efficiency measures − Strategies adopted to maintain editorial control − Perceptions and attitudes towards platform partners − In-house AI development potential HYPOTHETICAL RESULTS The following are hypothetical results based on existing research and trends synthesized from contemporary research and industry papers. Until the research is conducted, the following section is strictly hypothetical. Deepening Legal Battles Between Platforms and Publishers There is a rising trend in legal friction between news publishers and AI companies. 15 documented cases of copyright litigation are currently opened against AI developers and chip makers in the USA (Baker & Hostetler LLP, NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 68 n.d.). As previously mentioned, there is a unique, albeit conflicting, relationship between these companies. For instance, the NYT is directing its staff to use tools from OpenAI, the same company it is suing for copyright infringement (Benton, 2025). Until these cases begin to be resolved there is no clear path forward for news publishers. This paper will aim to uncover the strategies that news publishers will develop based on legal rulings. Economic Pressure as the Primary Driver of AI Adoption News publishers are oft en reliant on big tech firms that provide technical infrastructure and grants, which reinforce a growing dependence on external platforms (Sjøvaag, 2024). As the technology becomes more available publishers may transition to in-house solutions deploying their own AI tools evidenced by The Washington Post (Fischer, 2024). They may also partner with alternative AI developers, like universities, to build lowercost solutions (Adami, 2024). Hybrid AI-Editorial Workflows & Ethical Frameworks Successful AI ad opters will likely implement hybrid models, where AI will assist in tasks such as content personalization, preliminary interviews, data analysis, and content summaries while human journalists oversee quality control and verification (Beckett & Yaseen, 2023; Thäsler-Kordonouri, 2024). As more journalists use AI tools, ethical frameworks and AI usage policies will begin to emerge in leading newsrooms, which according to a Thomas Reuters Survey of 200 journalists around the globe, only 13% of newsrooms currently have (Radcliffe, 2024). NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 69 A Paradigm Shift in Journalism as a Profession Newsrooms may be more keen to hire journalists who are more tech-savvy and can use AI tools to develop stories, combining traditional journalistic knowledge with technical skills, as Danzon-Chambaud and Cornia (2024) state “...the adoption of a more abstract way of reasoning, which enables them [journalists] to master algorithmic production and to ensure that its outputs meet established professional standards and organisational n orms and practices.” (p.13). Lastly, journalists may begin to transition to self-branding, patenting their image and unique writing styles to protect themselves from encroaching AI systems that aim to replicate human journalists (Túñez-López et al., 2021). REFERENCES Adami, M. (2024). These Nordic newsrooms pioneered AI independently of Big Tech. Here’s what they learnt. Reuters Institute for the Study of Journalism. https://reutersinstitute.politics.ox.ac.uk/news/these-nordicnewsrooms-pioneered-ai-independently-big-tech-heres-what-theylearnt Adjin-Tettey, T. D., Muringa, T., Danso, S., & Zondi, S. (2024). The role of artificial intelligence in contemporary journalism practice in two African countries. Journalism and Media, 5, 846–860. https://doi.org/10.3390/journalmedia5030054 Ahmad, N., Haque, S., & Ibahrine, M. (2024). The news ecosystem in the age of AI: Evidence from the UAE. Journalism and Media, 5, 112–130. https://doi.org/10.1080/08838151.2023.2173197 Albizu-Rivas, I., Parratt-Fernández, S., & Mera-Fernández, M. (2024). Artificial intelligence in slow journalism: Journalists’ uses, perceptions, and attitudes. Journalism and Media, 5, 1836–1850. https://doi.org/10.3390/journalmedia5040111 NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 70 Beckett, C., & Yaseen, M. (2023). Generating Change: A global survey of what news organisations are doing with AI. Polis, JournalismAI, LSE. https://www.lse.ac.uk/media-and-communications/polis/JournalismAI Benton, J. (2025). The New York Times will let reporters use AI tools while its lawyers litigate AI tools. Nieman Journalism Lab. https://www.niemanlab.org/2025/02/the-new-york-times-will-letreporters-use-ai-tools-while-its-lawyers-litigate-ai-tools/ Braun, J. A., & Eklund, J. L. (2019). Fake news, real money: AdTech platforms, profit-driven hoaxes, and the business of journalism. Digital Journalism, 7(1), 1–20. https://doi.org/10.1080/21670811.2018.1556314 Carlson, M. (2015). The robotic reporter: Automated journalism and the redefinition of labor, compositional forms, and journalistic authority. Digital Journalism, 3(3), 416-431. https://doi.org/10.1080/21670811.2014.976412 Cools, H., & Diakopoulos, N. (2024). Uses of generative AI in the newsroom: Mapping journalists’ perceptions of perils and possibilities. Journalism and Media, 5, 45–63. https://doi.org/10.1080/17512786.2024.2394558 Danzon-Chambaud, S., & Cornia, A. (2024). The cultural capital you need to work with automated news: Not only “your beautiful piece of work,” but also “patterns that emerge.” Journalism. https://doi.org/10.1177/14648849241279579 de-Lima-Santos, M.-F., & Ceron, W. (2022). Artificial intelligence in news media: Current perceptions and future outlook. Journalism and Media, 3(1), 13-26. https://doi.org/10.3390/journalmedia3010002 Elliott, O. (2025, February). Representation of BBC News content in AI assistants. BBC Responsible AI Team. https://www.bbc.com/mediacentre/2025/bbc-research-shows-issueswith-answers-from-artificial-intelligence-assistants NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 71 Fischer, S. (2024, August 20). New Washington Post AI tool sifts massive data sets. Axios. https://www.axios.com/2024/08/20/washington-post-aitool-data Guion, L. A., Diehl, D. C., & McDonald, D. (2011). Conducting an in-depth interview. University of Florida, Institute of Food and Agricultural Sciences. https://doi.org/10.32473/edis-fy393-2011 Gutiérrez-Caneda, B., Vázquez-Herrero, J., & López-García, X. (2023). AI application in journalism: ChatGPT and the uses and risks of an emergent technology. Profesional de la información, 32(5), e320514. https://doi.org/10.3145/epi.2023.sep.14 Guzman, A. L., & Lewis, S. C. (2019). Artificial intelligence and communication: A human–machine communication research agenda. New Media & Society, 21(1), 21–36. https://doi.org/10.1177/146144481985869 Linden, C.-G. (2024). Decades of automation in the newsroom: Why are there still so many jobs in journalism? Journalism Studies, 25(3), 451–467. https://doi.org/10.1080/21670811.2016.1160791 Macalintal, A. (2024). Trustworthy AI: A comparative study of AI in journalism in Singapore and the Philippines. Digital Journalism, 12(2), 185–202. https://doi.org/10.18848/2470-9247/CGP/v09i01/99-123 Nechushtai, E. (2018). Could digital platforms capture the media through infrastructure? Journalism, 19(8), 1041–1058. https://doi.org/10.1177/1464884917725163 News Media Alliance. (2025). News Media Alliance announces industry lawsuit. News Media Alliance. https://www.newsmediaalliance.org/news-media-alliance-announcesindustry-lawsuit/ Robertson, K. (2024, April 30). Newspapers sue Microsoft and OpenAI over use of content. The New York Times. NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 72 https://www.nytimes.com/2024/04/30/business/media/newspaperssued-microsoft-openai.html Scheffauer, R., Gil de Zúñiga, H., & Correa, T. (2024). Algorithmic news versus non-algorithmic news: Towards a principle-based artificial intelligence (AI) theoretical framework of news media. Profesional de la Información, 33(1), e330009. https://doi.org/10.3145/epi.2024.0009 Soper, S. (2023, December 13). OpenAI, Axel Springer ink deal to use news content in ChatGPT. Bloomberg. https://www.bloomberg.com/news/articles/2023-12-13/openai-axelspringer-ink-deal-to-use-news-content-in-chatgpt Peña-Fernández, S., Meso-Ayerdi, K., Larrondo-Ureta, A., & Díaz-Noci, J. (2023). Without journalists, there is no journalism: the social dimension of generative artificial intelligence in the media. Profesional de la información, 32(2), e320227. https://doi.org/10.3145/epi.2023.mar.27 Pinto, M. C., & Barbosa, S. O. (2024). Artificial intelligence (AI) in Brazilian digital journalism: Historical context and innovative processes. Journalism and Media, 5, 325–341. https://doi.org/10.3390/journalmedia5010022 Radcliffe, D. (2024). Journalism in the AI Era: A TRF Insights survey. Thomson Reuters Foundation. https://www.trust.org/resource/airevolution-journalists-global-south/ Simon, F. M. (2022). Uneasy bedfellows: AI in the news, platform companies and the issue of journalistic autonomy. Digital Journalism, 10(10), 18321854. https://doi.org/10.1080/21670811.2022.2063150 Sjøvaag, H. (2024). The business of news in the AI economy. AI & Society, 39, 567–584. https://doi.org/10.1002/aaai.12172 Sjøvaag, H., & Owren, T. (2021). The non-substitutability of local news? Advertising and the decline of journalism’s umbrella market model. Nordicom Review, 42(1), 1–14. https://doi.org/10.2478/nor-2021-0001 NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 73 Thäsler-Kordonouri, S. (2024). What comes after the algorithm? An investigation of journalists’ post-editing of automated news text. Journalism Practice. https://doi.org/10.1080/17512786.2024.2404692 Túñez-López, J. M., Fieiras Ceide, C., & Vaz-Álvarez, M. (2021). Impact of artificial intelligence on journalism: Transformations in the company, products, contents, and professional profile. Communication & Society, 34(1), 177-193. https://doi.org/10.15581/003.34.1.177-193 Ufarte-Ruiz, M.-J., Murcia-Verdú, F.-J., & Túñez-López, J.-M. (2024). Use of artificial intelligence in synthetic media: First newsrooms without journalists. Digital Journalism, 12(1), 89–108. https://doi.org/10.3145/epi.2023.mar.03 Wu, S. (2024). Journalists as individual users of artificial intelligence: Examining journalists’ “value-motivated use” of ChatGPT and other AI tools within and without the newsroom. Journalism. https://doi.org/10.1177/14648849241303047 Wu, S., Tandoc, E. C., & Salmon, C. T. (2024). Journalism reconfigured: Assessing human–machine relations and the autonomous power of automation in news production. Digital Journalism, 12(2), 132–151. https://doi.org/10.1080/1461670X.2018.1521299 NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 80 RESEARCH ON DIGITAL COMMUNICATION NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 81 B ASQUE F EMINISM ON S OCIAL M EDIA Maider Eizmendi-Iraola [email protected] University of the Basque Country (UPV/EHU) Ainara Larrondo-Ureta ainara-larro[email protected] University of the Basque Country (UPV/EHU) Ainhoa Novo-Arbona [email protected] University of the Basque Country (UPV/EHU) Julen Orbegozo-Terradillos julen.orbego[email protected] University of the Basque Country (UPV/EHU) Simón Peña-Fernández simon.p[email protected]us University of the Basque Country (UPV/EHU) I NTRODUCTION The Internet and social networks have caused a seismic shift in the way we communicate. For this reason, the analysis of communicative relationships taking place in the digital realm has become a significant research area. The study of feminist claims and language normalization processes has become essential in the digital sphere; after all, technologies are not just a reflection of NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 82 society, they are society itself (Astudillo-Mendoza et al., 2023), and the power relations that dominate and are reproduced depend on how these networks are used. Social networks have provided feminism with opportunities for dialogue and collaboration (Ott, 2018), and along with other spaces and forums, women can actively participate in social and political discussions. Thus, social media can help raise awareness about feminist claims, but at the same time, they can serve as platforms for antifeminist messages and amplify patriarchal ideas (BanetWeiser, 2018). Additionally, from a linguistic perspective, the use of social networks has become crucial, as they can help revitalize languages and secure spaces that were previously unavailable. Finally, researchers have highlighted the similarities between the feminist movement and the Basque language movement, due to the way power rela tions operate within them (Agirre & Eskisabel, 2019). STATE OF THE ART Feminism has evolved in distinct waves throughout its history, each responding to the socio-political context of its time. Although there is still no full consensus, many theorists argue that feminism is currently immersed in its fourth wave. The fourth wave, closely linked to the #MeToo movement, focuses on issues of sexual harassment and violence, proving that feminism remains a vital force in modern society (Laudano, 2019). Otra de las características que se le atribuyen a esta ola es su tendencia global y, en ello, tienen gran impacto las redes sociales. In short, it has transformed the field of communication and the strategies of expression of social and political movements and, in the case of feminism, it has also offered them the opportunity to find a place that has not been found in traditional media. The feminist movement began a long time ago in an attempt to echo the main demands; in this way, it has united different media and tools, but the street has undoubtedly been the main area. The network NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 83 offers him a new place. Moreover, according to several experts, it is the basic element of the so-called fourth wave (Laudano, 2019; Munro, 2013). In the Basque Country, feminism was deeply intertwined with the struggles of the 1960s, particularly the cultural and political movements for independence, where women began to emerge as active participants (Epelde et al., 2015). During the 1970s, feminist movements gained autonomy, especially in Norther Basque Country, gradually spreading to the south despite the challenges posed by Francoist repression. The 1980s saw the formalisation of feminist institutions like Emakunde, and by the 1990s, feminism was increasingly integrated into academi a (Esteban, 2019). With regard to social networks, Basque feminism has responded in an organized manner to instances of digital misogyny through specific campaigns, such as the one surrounding 'La Manada' (Orbegozo, Morales & Larrondo, 2019). METHODOLOGY The research design and data analysis process has been developed employing a mixedmethods approach that combines both qualitative and quantitative techniques. The study focuses on Twitter, a public platform with a strong Basque-speaking community, as the primary space for examining feminist discourse in Basque. The sample selection involves identifying Twitter accounts and posts related to feminism and the Basque language, ensuring a diverse range of perspectives. The research employs thematic analysis t o explore the content of tweets, identifying recurring themes and key topics in the feminist discussions. Network analysis has also been used to map the relationships between key actors and to evaluate their influence on the overall discourse. This helps to identify central figures and their impact within the feminist online community. Additionally, engagement metrics such as likes, retweets, and replies are analysed to measure the reach and impact of feminist messages in Basque. NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 84 RESULTS The analysis of the network reveals that a relatively small group of accounts, often representing feminist organisations, activists, and scholars, play a central role in initiating discussions and influencing the wider conversation. These accounts are characterised by a high level of engagement, with their tweets being frequently retweeted and replied to. For instance, 20% of the analysed accounts generated over 50% of the engagement in the conversations, indicating their central role in the discourse. However, the study also identifies a broader group of users, including individuals who participate less frequently but still contribute to the diversity of perspectives within the conversation. This highlights the multiplicity of voices within the Basque feminist movement on social media. A significant portion of the messages focuses on topics such as gender equality and sexual violence. There is a notable engagement with broader societal issues, such as the fight against misogyny and patriarchy, with many messages explicitly addressing the need for structural social change. However, there is also a marked presence of messages relating to the intersection of feminism and the Basque language. Messages disseminated in Basque serve to share experiences of empowerment and sorority among women and feminist collectives. They also highlight the importance of the Basque language in transmitting feminist values and are used to promote or announce feminist events, initiatives, and actions of various kinds. Overall, the results show that social media, particularly Twitter, serves as a vital space for feminist activism in Basque, where discussions on gender and language intersect and contribute to ongoing societal debates. NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 85 D ISCUSSION Social media, especially Twitter, has become an essential platform for feminist activism, enabling a diverse range of voices to participate in public debates on gender equality and social justice. The research demonstrates that feminist conversations in Basque are not only focused on traditional feminist issues but also on the intersection of language and feminism. The study also concludes that a core group of influential accounts, including feminist organisations and activists, play a central role in driving the conversation, with high levels of engagement from their followers. However, the findings suggest that the broader feminist community on social media includes many less active participants who contribute to the diversity of discussions and perspectives. Furthermore, the research highlights the challe nges and opportunities that social media presents for feminist activism. While it offers a platform for visibility and mobilisation, it also exposes feminist discourse to antifeminist backlash and the reinforcement of patriarchal ideas. In conclusion, the study underscores the importance of understanding the dynamics of online feminist activism in the Basque context and its potential to foster social and linguistic change. REFERENCES Agirre-Dorronsoro, L., & Eskisabel-Larrañaga, I. (2016). Euskalgintza eta feminismoa: Identitateak berreraiki, demokrazia sendotu, boteretze kolektiboa bultzatu eta subalternitate eraldatzaile unibertsalak eraikitzeko proposamen bat. Bat: Soziolinguistika aldizkaria, 98, 11-22. Astudillo-Mendoza, P., Figueroa-Qiroz, V. & Astete-Martínez, C. (2023). Feminismo, comunidad de mujeres y redes sociales online: Etnografía NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 86 digital en un grupo de Facebook. Athenea Digital, 23(3), e3321. https://doi.org/10.5565/rev/athenea.3321 Banet-Weiser, S. (2018) Postfeminism and popular feminism. Feminist Media Histories, 4(2). pp. 152-156. http://doi.org/10.1525/fmh.2018.4.2.152 Eizmendi-Iraola, M., Larrondo-Ureta, A., Novo-Arbona, A., OrbegozoTerradillos, J., & Peña-Fernández, S. (2024). Feminismoaren ahotsa euskaraz eta sarean. Gaiak, eztabaidak eta ikusgarritasuna. Eusko Jaurlaritza. Epelde, E., Aranguren, M., & Retolaza, I. (2015). Gure genealogía feminista. Txalaparta. Esteban, M. L. (2019). El feminismo y las transformaciones en la política. Bellaterra. Laudano, C. (2019). #Ni una menos en Argentina : Activismo digital y estrategias feministas contra la violencia hacia las mujeres. In G. Nathansohn y F. Rovetto (Org.), Internet e feminismos : olhares sobre violências sexistas desde América Latina. (pp. 149-173). EDUFBA. https://www.memoria.fahce.unlp.edu.ar/libros/pm.3711/pm.3711.pdf Munro, E. (2013). Feminism: A Fourth Wave? Political Insight, 4(2), 22-25. https://doi-org.ehu.idm.oclc.org/10.1111/2041-9066.12021 Orbegozo-Terradillos, J., Morales-i-Grass, J., & Larrondo-Ureta, A. (2019). Feminismos indignados ante la justicia: la conversación digital en el caso de La Manada. IC Revista Científica De Información Y Comunicación, 16. https://icjournal-ojs.org/index.php/IC-Journal/article/view/466 Ott, K. (2018). Social Media and Feminist Values: Aligned or Maligned? Frontiers, 39(1), 93-111. http://dx.doi.org/10.5250/fronjwomestud.39.1.0093 NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 87 M EDIA D ESENSITIZATION IN C OVERAGE OF M ASS AND S CHOOL SHOOTINGS: STATE OF THE ART AND FUTURE RESEARCH PROPOSAL Kyle Leaver mailto:[email protected] Pompeu Fabra University (UPF) I NTRODUCTION Mass shootings and school shootings have increased in frequency in the United States, leading to intense and repetitive media coverage (Gun Violence Archive, 2023; Schildkraut & Muschert, 2014). While journalists play a critical role in documenting these tragedies, they also experience repeated exposure to graphic content, survivor testimonies, and crime scenes. Thus, raising concerns about emotional desensitization and traumarelated stress (Feinstein et al., 2014; Simpson & Coté, 2006). M edia desensitization, the process by which repeated exposure to trauma reduces emotional responsiveness, is welldocumented in media effects research (Lazarsfeld & Merton, 1948; Sparks & Sparks, 2000). Most studies have focused on public desensitization, showing that continuous exposure to violent media lowers empathic concern and emotional arousal in audiences (Oliver & Bartsch, 2010; Potter, 2006). However, journalists are uniquely positioned as both witnesses and reporters, facing first-hand exposure to real-world violence rather than mediated portrayals (Simpson & Boggs, 1999; Greenberg et al., 2009). The cumulative impact of this exposure on journalists’ mental health, ethical decisionmaking, and reporting styles remains underexplored (Newman et al., 2003; Leaver, 2023). NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 88 This study investigates how repeated mass shooting coverage affects journalists’ psychological wellbeing, ethical dilemmas, and emotional engagement. Using media framing theory (Entman, 1993), trauma journalism research, and empirical data from mass shooting archives, this paper argues for traumainformed newsroom policies that safeguard journalist mental health while maintaining ethical reporting standards (McGinty et al., 2019; Meindl & Ivy, 2017). STATE OF THE ART: UNDERSTANDING MEDIA DESENSITIZATION The Role of the Media in Mass Shooting Coverage Journalists serve as primary information providers in the aftermath of mass shootings, shaping public discourse and policy debates (Duwe, 2007; Altheide, 2009). Studies show that media coverage of mass shootings follows recurring narrative structures, typically focusing on: − Shooter profiling – Investigating perpetrators' backgrounds, motives, and mental health histories (Schildkraut & Muschert, 2014; Silva & Capellan, 2019). − Gun policy debates – Discussions on firearm regulation and legal loopholes (McGinty et al., 2019). − Victim-centered storytelling – Humanizing tragedies through survivor and family narratives (Patterson, 2013). − Law enforcement response – Evaluating police intervention strategies (Fox & DeLateur, 2014). While these elements provide essential context, research suggests that repetitive media framing can normalize mass violence, reducing its emotional impact on both journalists and audiences (Kissner, 2016; Leaver, 2023). Additionally, extensive coverage of perpetrators has been linked to copycat shootings, a phenomenon known as the contagion effect (Meindl & Ivy, 2017; Towers et al., 2015). NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 89 Psychological Impact on Journalists Journalists covering mass shootings experience long-term psychological stress, with symptoms similar to those observed in war correspondents and trauma workers (Feinstein et al., 2014; Greenberg et al., 2009). Studies have identified several adverse effects, including: − Secondary Traumatic Stress (STS) – PTSD-like symptoms from repeated exposure to traumatic content (Simpson & Coté, 2006). − Emotional numbing – Decreased emotional engagement and empathy over time (Newman et al., 2003). − Moral injury – Cognitive dissonance stemming from ethical dilemmas in coverage (Pyevich et al., 2003). Greenberg et al. (2009) found that journalists often suppress their emotional responses to maintain objectivity, but this leads to chronic emotional exhaustion and detachment. Moreover, Leaver (2023) suggests that newsroom culture prioritizes productivity over psychological wellbeing, discouraging journalists from seeking support for trauma-related stress. Ethical Dilemmas in Reporting Mass Shootings Journalists covering mass shootings must navigate difficult ethical decisions, including: − Shooter glorification – Extensive focus on perpetrators may unintentionally promote notoriety (Meindl & Ivy, 2017; Schildkraut & Elsass, 2016). − Use of graphic imagery – Publishing violent visuals can re-traumatize victims' families while desensitizing audiences (Patterson, 2013). − Sensationalism vs. Objectivity – The pressure to attract viewership may lead to emotionally exploitative reporting (McGinty et al., 2019). NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 96 Simpson, R., & Boggs, J. (1999). An exploratory study of traumatic stress among newspaper journalists. Journalism & Communication Monographs, 1(1), 1-26. https://doi.org/10.1177/152263799900100102 Simpson, R., & Coté, W. (2006). Covering violence: A guide to ethical reporting. Columbia University Press. Sparks, G. G., & Sparks, C. W. (2000). Violence, mayhem, and horror: The communicative effects of mass media violence. Media Effects: Advances in Theory and Research, 269-285. Towers, S., Gomez-Lievano, A., Khan, M., Mubayi, A., & Castillo-Chavez, C. (2015). Contagion in mass killings and school shootings. PLoS ONE, 10(7), e0117259. https://doi.org/10.1371/journal.pone.0117259 NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 97 B REAST C ANCER ON T IK T OK : E MOTIONAL S UPPORT , MISINFORMATION, AND THE CHALLENGE OF RELIABLE HEALTH COMMUNICATION Jesús Ángel Pérez-Dasilva jesusangel.p[email protected]s University of the Basque Country (UPV/EHU) María Ganzabal-Learreta [email protected] University of the Basque Country (UPV/EHU) Urko Peña-Alonso [email protected] University of the Basque Country (UPV/EHU) I NTRODUCTION Breast cancer is a complex disease influenced by multiple risk factors, including genetic predisposition, hormonal changes, and lifestyle factors such as diet, alcohol consumption, smoking, and obesity (Martínez Basurto et al., 2014). Beyond its medical implications, a breast cancer diagnosis carries a profound emotional and psychological burden, affecting patients' well-being on multiple levels. In the face of these challenges, many individuals turn to digital communities for emotional support, practical guidance, and a sense of belonging (Abt Sacks et al., 2013). Online interactions often foster empathy and gratitude, which Han NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 98 et al. (2011) identify as key mechanisms of emotional support. Additionally, studies highlight encouragement, religious faith, and direct requests for help as common forms of engagement within these spaces (Kim et al., 2012). Social media platforms play a significant role in supporting breast cancer patients. Research by Attai et al. (2015) suggests that participating in Twitter discussions reduces anxiety and increases disease awareness. Similarly, Pluta (2022) argues that Instagram fosters positive emotions, selfacceptance, and shared experiences of coping. On TikTok, patients frequently share survival stories, educational content, and personal narratives, which help build community and awareness (Pluta & Siuda, 2022; Wellman et al., 2022). However, not all content is beneficial. Johnson et al. (2022) report that over 30% of cancerrelated content on Facebook contains misinformation. Similar concerns arise on TikTok, where Siva et al. (2022), and Liu et al. (2024) found that approximately 40% of breast cancer videos are created by non-medical users and often contain unreli able information. This is alarming, as cancer patients are particularly susceptible to misinformation and pseudoscientific treatments. Van Prooijen (2019) explains that individuals facing severe illnesses may seek out "magical" solutions for hope, while Čavojová et al. (2023) found that women with cancer are significantly more likely to believe in pseudoscientific claims than those without the disease. Interestingly, medical professionals and health organizations create higherquality content, yet their videos receive fewer views. This raises concerns about the need for healthcare professionals to adapt their communication strategies to reach larger audiences without compromising accuracy (Pérez & Ganzabal, 2023). Studies suggest that while medical content is often too technical, personal storytelling resonates more strongly with viewers, making it a more effective engagement tool. NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 99 A NALYSIS OF B REAST C ANCER C ONTENT ON T IK T OK To examine the content, determine the profiles of users posting about this carcinoma, and identify the most recurring topics, we collected Spanishlanguage TikTok videos tagged with "breast cancer" between October 2022 and 2023 using Zeeschuimer software (Pérez et al., 2025). These posts were sorted by engagement, and the top 270 were analyzed, focusing on aspects such as the role of the protagonist, age, emotional state, or type of content published. The study categorized TikTok accounts discussing breast cancer into three main groups: 1. General content creators (49%), who occasionally address breast cancer. 2. Accounts exclusively dedicated to breast cancer (29%), often run by patients documenting their experiences. 3. Cancerfocused accounts (22%), covering breast cancer as part of broader oncological discussions. Most videos are short, wi th 58.42% lasting under a minute, though longer formats are becoming more common (20.22% around two minutes, 21.3% over two minutes). Surprisingly, only 10.98% of videos are created by medical professionals, and no active contributions from official health organizations, NGOs, or foundations were found. Women feature in 60% of the videos, primarily as patients sharing personal stories. Men appear in 27.78% of videos, usually discussing prevention rather than personal experiences. A small percentage (2.2%) includes actors or individuals without direct experience with the disease. Regarding age demographics, the largest group (35.56%) is between 25 and 40 years old, followed by 40–65-yearolds (13.33%). Younger patients tend to NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 100 emphasize hope and resilience, whereas older patients focus more on disease progression and treatment impact. Notably, only 1% of videos feature women over 65. A significant portion of the content (31.87%) portrays breast cancer as a life-ordeath battle, while 28.57% highlight the physical consequences of treatment, such as mastectomies. In contrast, only 10.99% of videos focus on survivors who have overcome the disease, though many still discuss lingering effects like hair loss or lymphedema. MISINFORMATION ABOUT BREAST CANCER ON TIKTOK Misinformation is the most prevalent theme in breast cancer-related TikTok videos (23.08%). Unverified treatments and misleading claims— such as Ludwig Johnson’s promotion of carrots as a cancer cure—are widely shared. Using criteria from the Spanish Association Against Cancer (AECC), researchers identified common misinformation tactics, including: − Sensationalist titles − Distorted interpretations of scientific research − Oversimplification of complex medical concepts − Omission of critical details Other m ajor themes include discussions about conventional treatments (19.78%), general breast cancer information and prevention (18.13%), messages of hope (12.09%), and community support (10.99%). Additionally, 7.69% of videos critique healthcare access disparities. CONCLUSIONS The study confirms that TikTok serves as a platform where breast cancer patients share their experiences and connect with others. The most viral videos are those that evoke strong emotions—32% depict the struggle against cancer, NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 101 while 28.6% focus on the physical consequences of treatment. These narratives help humanize the disease and foster empathy (Martínez & Arribas, 2023). However, misinformation remains a significant issue, with 23.08% of videos promoting unverified claims. The prevalence of pseudoscientific content poses a public health risk, as it can influence patients' decisionmaking and encourage reliance on ineffective treatments (Porroche, 2017). TikTok itself is not the root cause of misinformation; rather, it is the delibera te use of sensationalist strategies that attract viewers at the expense of accuracy. In conclusion, while TikTok provides valuable emotional support and awareness, stronger efforts are needed to combat misinformation and enhance the credibility of breast cancer content. Encouraging medical professionals to adopt more engaging storytelling techniques could help bridge the gap between reliable information and audience engagement. REFERENCES Abt Sacks, A., Pablo Hernando, S., Serrano Aguilar, P., Fernández Vega, E., & Martín Fernández, R. (2013). Necesidades de información y uso de Internet en pacientes con cáncer de mama en España. Gaceta Sanitaria, 27(3), 241-247. https://doi.org/10.1016/j.gaceta.2012.06.014 Attai, D. J, Cowher, M. S., Al-Hamadani, M., Schoger, J. M., Staley, A. C., &Landercasper, J. (2015). Twitter Social Media is an Effective Tool for Breast Cancer Patient Education and Support: Patient-Reported Outcomes by Survey. Journal of Medical Internet Research,17(7), e188. https://doi.org/10.2196/jmir.4721 Čavojová, V., Kaššaiová, Z., Šrol, J., &Mikušková, E. B. (2023). Thinking magically or thinking scientifically: Cognitive and belief predictors of complementary and alternative medicine use in women with and NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 102 without cancer diagnosis. Current Psychology, 43, 7667–7678. https://doi.org/10.1007/s12144-023-04911-8 Han, H. S., Reis, I. M., Zhao, W., Kuroi. K., Toi, M., Suzuki, E., Syme, R., Chow, L., Yip, A.Y., &Glück, S. (2011). Racial differences in acute toxicities of neoadjuvant or adjuvant chemotherapy in patients with early-stage breast cancer. European Journal of Cancer, 47(17), 2537-45. https://doi.org/10.1016/j.ejca.2011.06.027 Johnson, S. B., Parsons, M., Dorff, T., Moran, M. S., Ward, J. H., Cohen, S. A., Akerley, W., Bauman, J., Hubbard, J., Spratt, D. E., Bylund, C. L., Swire-Thompson, B., Onega, T., Scherer, L. D., Tward, J., & Fagerlin, A. (2022). Cancer misinformation and harmful information on Facebook and other social media: A brief report. JNCI: Journal of the National Cancer Institute, 114(7), 1036-1039. https://doi.org/10.1093/jnci/djab141 Kim, S. C., Shah, D. V., Namkoong, K., McTavish, F. M., & Gustafson, D. H. (2013). Predictors of online health information seeking among women with breast cancer: The role of social support perception and emotional well-being. Journal of Computer-Mediated Communication, 18(2), 212-232. https://doi.org/10.1111/jcc4.12002 Liu, H., Peng, J., Li, L., Deng, A., Huang, X., Yin, G., Ming, J., Luo, H. & Liang, Y. (2024) Assessment of the reliability and quality of breast cancer related videos on TikTok and Bilibili: cross-sectional study in China. Frontiers in Public Health, 11, 1296386. https://doi.org/10.1007/s11695019-03738-2 Martínez-Sanz, R. &Arribas-Urrutia, A. (2023). Blood donors wanted: narrative innovation on TikTok to enable mobilization. Profesional de la información, 32(3). https://doi.org/10.3145/epi.2023.may.05 Pérez-Dasilva, J. Á. & Ganzabal-Learreta, M. (2023). New platforms for information: Choreographies to tell the news on TikTok. In S. PeñaFernández & K. Meso-Ayerdi (Eds.), News in the Hybrid Media system (pp. 83-92). Universidad del País Vasco. NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 103 Pérez-Dasilva, J. Á., Ganzabal-Learreta, M., Meso-Ayerdi, K., Peña-Alonso, U., & Mendiguren-Galdospin, T. (2025). Cáncer de mama en el bazar de TikTok: Testimonios frente a charlatanería. Cuadernos.Info, 60, 121– 142. https://doi.org/10.7764/cdi.60.82510 Pluta, M. (2022). Online Self-Disclosure and Social Sharing of Emotions of Women with Breast Cancer Using Instagram–Qualitative Conventional Content Analysis. Chronic Illness, 18(4), 834-848. https://doi.org/10.1177/17423953211039778 Pluta, M. & Siuda, P. (2022). Cancer on TikTok–evaluating positive culture and online selfdisclosure using directed content analysis and in-depth interviews. AoIR Selected Papers of Internet Research. https://doi.org/10.5210/spir.v2022i0.13070 Porroche-Escudero, A. (2017). Problematizando la desinformación en las campañas de concienciación sobre el cáncer de mama. Gaceta Sanitaria, 31(3), 250-252. https://doi.org/10.1016/j.gaceta.2016.11.003 Siva, N., Koirala, M., Raiker, R., Waris, S., Pakhchanian, H., & Puckett, Y. (2022). Evaluation of trends in breast cancer-related content on TikTok. Poster session presented at the West Virginia University School of Medicine, Morgantown, WV; Kansas City University College of Osteopathic Medicine, Kansas City, MO; George Washington University School of Medicine and Health Sciences, Washington. Van Prooijien, J. W. (2019). An Existential Threat Model of Conspiracy Theories. European Psicologist, 25(1). https://doi.org/10.1027/10169040/a000381 NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 104 C RITICAL MEDIA LITERACY AND MIGRATION IN HIGHER EDUCATION: A CRITICAL DIALOGIC APPROACH Mayte Santos-Albardia [email protected] University of the Basque Country (UPV/EHU) I NTRODUCTION In the current information context, media discourses on migration tend to reinforce narratives that distort reality and contribute to the construction of an “otherness” marked by rejection and stigmatization (Creus and Martins, 2024). Although the migrant population represents a moderate percentage in Spain, social perceptions tend to overestimate their presence and impact. Previous studies have shown that disinformation and the repetition of media stereotypes have direct effects on the acceptance of inclusion policies and the integration of migrants into local communities (Ikuspegi, 2023). Critical media literacy has been identified as a key tool to foster understanding of media discourses and provide citizens with tools to analyze and question the messages they receive on a daily basis (Buitrago Alonso et al., 2017). However, its implementation at the university level remains limited, partly owing to the lack of teacher training and the scarcity of public policies for its promotion in a structured way (Bezanilla et al., 2021). This study proposes a methodology that combines expert interviews, questionnaires, and focus groups with dialogic strategies in the classroom to evaluate the effectiveness of critical media literacy in university education. The aim is to understand how such training in critical educommunication can NEWSNET #5 THE IMPACT OF ARTIFICIAL INTELLIGENCE ON SOCIAL COMMUNICATION 105 contribute to the development of a more informed perception regarding migration and the effects of the associated media discourses. METHODOLOGY T he general objective of this study is to discuss the potential of educommunication from a critical dialogic viewpoint, as a solution to provide tools that can generate critical thinking among citizens. The specific objectives were: O1. To identify students’ perception about migration and the role of the media in its construction. O2. To implement a methodology based on critical dialog in the classroom to foster critical reading of communications on migration. O3. To assess the impact of critical media literacy as a potential crosscutting competency in higher education. The research adopts a qualitative approach, combining seven in-depth interviews held with experts in critical media literacy and diagnostic questionnaires and focus groups involving 141 Education and Communication students at the UPV/EHU. NVIVO software was used for data analysis, allowing accurate coding, as well as content and critical discourse analyses (Gee, 2004). Indepth interviews were conducted with seven experts in the field of ed ucommunication and critical media literacy from different regions, including North America, Latin America, and Spain. Their discourses were analyzed on the basis of conceptual differences in the field of educommunication, the need to strengthen critical media literacy among both teachers and students, and the role of public policies in its implementation.