Back to All Events

CRAI-CIS Seminar: The value of structural features for cross-language prominent content analysis

  • Aalto University Konemiehentie 2 (map)

Critical AI & Crisis Interrogatives (CRAI-CIS) Seminar

Monthly dialogues and critical perspectives on artificial intelligence, Human Computer Interaction (HCI), participatory design, and crisis-related research for societal impact.

The CRAI-CIS Seminars engage emerging work across critical AI, Human Computer Interaction (HCI), participatory design, and crisis-related research. The seminars seek to invoke dialogues on how computational, human-centred, and social sciences perspectives can offer new insights and methods for inclusive approaches and critical inquiry with societal impact.

Each event features invited speakers who share distinct perspectives, ongoing research, methods, and challenges for future work in a 45 minute talk, followed by Q&A and space for mingling and networking. The talks will be recorded for open access in the future.

Lear more about the seminar series →


Asta Zelenkauskaite: The value of structural features for cross-language prominent content analysis

Time: November 27, 2024, 16:00–17:30
Venue: Event is hybrid. Participants may attend in person (T2, Computer Science Building, Aalto University) or online on Zoom.

Abstract: 

While link analysis is crucial for understanding online content dynamics, we argue that more focus should be placed on extracting insights that can be meaningful for cross-language comparative contexts. In this talk, we explore how analyzing shared links on mobile platforms where people may exchange anonymous information, such as Telegram, can reveal popular content and emerging themes. By examining duplicate links in cross-language fringe online spaces, we demonstrate how a significant portion of shared links highlights the spread of conspiracy theories. Based on Telegram examples, this work will showcase a list of techniques that can be used to uncover patterns of duplicates by using context- and domain-specific approaches. This analysis, which uses both bottom-up and top-down approaches, uncovers unique and common patterns in different language networks, providing a deeper understanding of content dissemination, networks of spread, and temporal dimensions.

Speaker Bio: 

Dr. Asta Zelenkauskaite, is professor of Communication at Drexel University. Her work focuses on the societal challenges of information mistrust and post-truth, and how such inauthentic information can be uncovered. She employs multidisciplinary approaches drawn from social science traditions, communication, information science, and linguistics, studying these phenomena through both macro and micro methods. She is the author of Creating Chaos Online: Disinformation and Subverted Post-Publics (University of Michigan Press). Her work has also been published in New Media & Society, Social Media + Society, Journal of Communication, Convergence, and First Monday, among others. She co-chairs the IAMCR Audience section and is affiliated with the Science, Technology, and Society and Information Science departments at Drexel, as well as with Vilnius Tech in Lithuania.