New CDT Report: Moderating Quechua Content on Social Media
Hi everyone, We are excited to announce the publication of a new research report from the Center for Democracy & Technology (CDT), "*Moderating Quechua Content on Social Media.* <https://cdt.org/insights/moderating-quechua-content-on-social-media/>"We explore how social media platforms conduct analysis of user generated content in Quechua, as well as processes for enforcement and appeals of moderation decisions. In addition, many platforms use large language models (LLMs) and other automated tools to support their content moderation systems. We examine how such technologies are incorporated into trust and safety systems that handle content in Quechua. This study is part of CDT's research series on the challenges of moderating low-resource languages in the Global South (see our related research on Maghrebi Arabic <https://cdt.org/insights/moderating-maghrebi-arabic-content-on-social-media/>, Kiswahili <https://cdt.org/insights/moderating-kiswahili-content-on-social-media/>, and Tamil <https://cdt.org/insights/moderating-tamil-content-on-social-media/>). Using a mixed-method approach that includes an online survey of Quechua social media users, interviews with content moderators, influencers, linguistic activists, and digital rights advocates from Ecuador, Peru and Bolivia, we well as extended discussions with NLP researchers and LLM developers working on language technologies in Quechua, we found that: * Quechua social media users most frequently use Facebook, but most use multiple platforms, with a combination of Quechua and Spanish. * The motivation for some Quechua social media users to post online is to share their language and culture with each other and the world.Several of our interviewees affirmed that this form of linguistic activism and resistance, often done in Quechua. * Quechua social media users are confronted with multiple problems on platforms, including what they perceive as unjust content removals and algorithmic discrimination.Their own attempts at testing the platforms’ algorithms point to examples of the differential treatment between Quechua and Spanish, where moderation of the latter appeared to be more effective. These problems are driven in part by the limitations of the technologies (automated tools, LLMs, etc.) used by social media platforms. * There is a significant problem of harassment targeted at those who post in Quechua.As some interviewees recognized, part of this problem is also gendered, with women who identify as Quechua online subject to online gender based violence. * Automated content moderation technologies and trust and safety processes appear to fail Quechua social media users in several ways.Experts suggested that LLMs are not yet ready to be applied to content moderation in Quechua, given the lack of training data and other limitations. In addition, content moderation teams lack Quechua speakers and analysts, and the content policies used were essentially copied from English and Spanish and applied to Quechua content, leading to indecision, delays, and the non-removal of harmful content. We conclude with recommendations to create trust and safety systems that support language equity while respecting the rights and aspirations of those who post in indigenous languages such as Quechua. You can read and share threads on Bsky <https://bsky.app/profile/cdt.org/post/3lsjoczbofa2x>, LinkedIn <https://www.linkedin.com/posts/center-for-democracy-%26-technology_how-do-content-moderation-systems-treat-indigenous-activity-7344057259594821633-TxRH?utm_source=share&utm_medium=member_desktop&rcm=ACoAABmRVjQBbSXIuPC3_K9dGpdU_vv3AzZW62k>and Mastodon <https://techpolicy.social/@CenDemTech/114750893089040642>, and you can access the full report here <https://cdt.org/insights/moderating-quechua-content-on-social-media/>for more details. We will soon publish a video summary of the report in Quechua and a Spanish translation (using the same link). Feel free to share within your networks. Your feedback and insights are important to us as we continue exploring the complexities of content moderation in other regions in the Global South. Best regards,Dhanaraj -- *Dhanaraj Thakur* (he/him) | Research Director Center for Democracy & Technology |*cdt.org <https://cdt.org/>* **dthakur@cdt.org | **+1 202 407 8849
Seventh International Conference on Real-Time Intelligent Systems (RTIS 2025) Macau Millennium College, Macau October 13-14, 2025 Posts and Telecommunications Institute of Technology November 05-06, 2025 Ho Chi Minh. Vietnam https://www.socio.org.uk/rtis Authors can participate in Macau or in Vietnam. The International Conference on Real-time Intelligent Systems (RTIS) has travelled from Beijing, China (2016), to Macau. The seventh edition will take place at the University of Macau, Macau and at the Posts and Telecommunications Institute of Technology, Ho Chi Minh, in Vietnam. Over the past few years, real-time intelligent computing has revolutionised the human lifestyle. Research on real-time intelligent systems is multi-disciplinary, exploiting concepts from diverse areas such as big data processing, computational intelligence, location-based services, recommendation systems, and multimedia processing. In today’s highly dynamic environment, analysing data in real-time is necessary to understand how systems process data, reason the outputs, and anticipate trends in intelligent computing. To this end, this conference will serve as a platform to manifest the ongoing research in the field. Thus, RTIS welcomes theoretically grounded, methodologically sound papers that address aspects related to the following topics. CALL FOR PAPERS (Virtual/Physical) Artificial Intelligence and Data Mining Streaming data, streaming engines Trace-based intelligent real-time services Adaptive vision algorithms Location-based services Intelligent Robotic Systems Collaborative Intelligence Data capture in real-time Data quality and cleansing Intelligent Data Analysis Intelligent Database Systems Knowledge representation and reasoning Intelligent information fusion Large Language Models, cognitive methods, sequential inference, data mining, pattern/behavioral analysis Big Data systems and applications for high-velocity data Intelligent Information Systems Privacy and security in Intelligence Software Engineering Solutions Intelligent Soft Computing Real-time multiprocessor systems Internet of Things Architectures for Intelligence Real-time distributed coding Smart services and platforms Real-time modelling user information needs Wireless Communication Real-time intelligent communication Real-time intelligent network solutions Mobile Smart Systems Broadband Intelligence Cloud Computing and Intelligence Collaborative Intelligence Analysis in domains such as energy, sensors Decision support systems in real-time Multi-agent Intelligent Systems Multilingual information access Recommendation systems Real-time intelligent alert systems Real-time remote access systems Intelligent Transportation Systems Autonomous systems (incl. autonomous vehicles and drones) Distributed systems Cloud/edge computing/fusion Defence/security, robotics, aerospace, intelligent transportation Mining/Manufacturing Environmental monitoring Critical Real-time Applications Real-time noise removal systems Event-driven analytics Intelligent Fuzzy Systems Machine translation in real-time OLAP for real-time decision support Crowdsourcing and crowd intelligence AI in Agriculture Smart Agriculture Systems User experience in Real-Time Systems Anomaly Detection Predictive Maintenance Image and Video Recognition Natural Language Processing Vulnerability Management Healthcare Customer Analytics Submission, proceedings Papers must be submitted online through OpenConf. Author instructions and LaTex2e (preferred) and Word macro files are available on the submission page. Submitted papers should be at most 14 pages (long papers) and 8 pages (short ones), including figures, tables and references (in the Springer template). Authors of accepted papers are required to transfer their copyrights. For a paper to appear in the proceedings, at least one of the authors MUST register for the conference by the camera-ready submission deadline with a full registration. Springer’s Lecture Notes in Networks and Systems (LNNS) (https://www.springer.com/series/15179) will publish the accepted papers and be indexed in SCOPUS, EI Compendex, INSPEC, WTI Frankfurt eG, zbMATH, and SCImago. All the papers published in the series are submitted for consideration in the Web of Science. Important Dates Submission of Papers: August 25, 2025 Notification of Acceptance/Rejection: September 20, 2025 Camera-ready: October 05, 2025 Registration: October 05, 2025 Conference Dates: October 13-14, 2025 General Chair Simon Fong, University of Macau, Macau Program Chairs Pit Pichappan, Digital Information Research Labs, India Zehan Tan, Macau Millennium College, Macau Program Co-chairs Kenneth Chan, Macau Millennium College, Macau Duong Van Hieu, Tien Giang University, Vietnam Ricardo Rodriguez-Jorge, Technological Centre Ceit, Spain Dion Goh, Nanyang Technological University, Singapore Contact: stm@socio.org.uk
Second International Workshop on Data Analytics in Biomedicine held in conjunction with DDP2025 (IEEE) Luton, UK 18-20 August 2025 https://socio.org.uk/ddp/workshopinter.php CALL FOR PAPERS The exponential growth of data generated from various textual sources presents both a challenge and a huge opportunity. A key challenge lies in effectively managing and extracting valuable insights from this vast amount of unstructured and heterogeneous data. To address this issue, advanced data analytics techniques, ranging from data and text mining to semantic network analysis and recent advancements in large language models (LLMs), have become indispensable tools for researchers and practitioners. This is particularly relevant in the realm of biomedicine, where text mining has shown the ability to enable researchers to uncover hidden patterns, trends, and associations that would otherwise remain buried in the vast amount of health-related textual data, as research articles, clinical reports, and electronic health records (EHRs). On the other hand, semantic network analysis, which focuses on understanding the structure and dynamics of networks formed by entities and their interconnections derived from text mining processes, can facilitate a deeper understanding of the complex interrelationships within biomedical data. By analyzing properties like centrality, modularity, and community structures, researchers can identify key nodes and critical pathways in biological networks, predict disease associations, and explore the functional organization of biological systems. Integrating text mining, semantic network analysis, and large language models offers a powerful approach to enhance the ability to generate new hypotheses and insights and support the development of more effective diagnostics, treatments, and interventions. The workshop represents an opportunity to explore the latest advancements in data analytics and text mining in biomedicine. Attendees will gain insights into developing more interpretable models, handling large-scale biomedical datasets, and implementing scalable solutions for real-world healthcare applications. Moreover, the workshop is highly relevant because it has the potential to significantly improve the safety, effectiveness, and efficiency of biomedical interventions through advanced data analytics. TOPIC OF INTEREST We invite submissions on a wide range of topics, including but not limited to: • Novel techniques and measures for assessing textual data quality and handling data integration. • Advanced text-mining techniques for biomedical data • Construction and analysis of semantic networks in biomedicine • Case studies on integrated text mining and semantic network analysis • Applications of LLMs in biomedicine • Data analytics in precision medicine • Text-driven approaches to drug discovery • Interpretable or scalable data analytics approaches • Application of data analytics and network science in narrative medicine • Computational methods for disease modeling and prediction • Ethical considerations in biomedical data analytics • Multimodal biomedical data analytics • Future trends and challenges in biomedical data analytics PROGRAM The workshop will take place on (To Be Announced). The program is not available yet. PAPER SUBMISSION – REGISTRATION AND PUBLICATION Please refer to socio.org.uk/ddp/ IMPORTANT DATES Submission of Papers: July 15, 2025 Notification of Acceptance/Rejection: August 10, 2025 Camera-ready: August 31, 2025 Registration: August 31, 2025 Conference Dates: August 18-20, 2025 Post-Conference Proceedings Release: November 30, 2025 WORKSHOP ORGANIZERS Chiara Zucco, University Magna Graecia of Catanzaro, Italy Mario Cannataro, University Magna Graecia of Catanzaro, Italy Marianna Milano, University Magna Graecia of Catanzaro, Italy PROGRAM COMMITTEE (TO BE CONFIRMED) Marzia Settino, University of Calabria, Italy Mario Cannataro, University Magna Graecia of Catanzaro, Italy Maria Chiara Martinis, University Magna Graecia of Catanzaro, Italy Giuseppe Agapito, University Magna Graecia of Catanzaro, Italy Pietro Cinaglia, University Magna Graecia of Catanzaro, Italy Ilaria Lazzaro, University Magna Graecia of Catanzaro, Italy Contact: ddp@socio.org.uk
participants (2)
-
Dhanaraj Thakur -
ijwa＠dline.info