AI, ethnographic field devices, and trying to understand visual war propaganda
The use of AI in research is becoming a heated philosophical debate among researchers, academics and institutions, with both utopian and dystopian visions of what this means for established conventions of research and the future of knowledge. These debates are important - but we will not talk about them in this presentation. Instead, we want to focus on a much more mundane and practical use case where we have been experimenting with different AI-assisted workflows to augment and, at times, shape the type of research questions we are able to address. We illustrate this problem by focusing on our ongoing efforts to understand information suppression and propaganda during the Tigray War (2020-2022) in Ethiopia. Through discussing our effort to analyse and understand over 500,000 war images shared on Twitter during the war, we explore - and sometimes confess - how we somewhat clumsily stumbled upon a provisional workflow for using different AI tools (such as LLM-based coding assistants, Chinese vision-language models, and shared custom dashboards for data validation and exploration) to help solve our specific research problem. On a more methodological level, in our talk, we highlight what we call experimental "field devices”, provisional research tools that can be one way to augment qualitative approaches to research, sometimes with the help of proliferating AI agents.
Matti Pohjonen is a University Researcher at the Helsinki Institute for Social Sciences and Humanities and Research fellow in AI in Social Research on Diversity for DIVSOL, at the University of Helsinki. An anthropologist by background, his current research focuses on digital politics in comparative global contexts, with a focus on generative AI and large language models (LLMs) as emerging forms of global knowledge production. He also currently co-leads an EU Horizon Europe project on authoritarian information suppression, with a focus on Ethiopia and its diaspora in Europe.
Amanuel Tesfaye is a doctoral researcher at the University of Helsinki, researching propaganda, strategic narratives, and authoritarian adaptation in violent conflict contexts, with a focus on Ethiopia. He has researched social movements, nationalist mobilization, and the political economy of development in Ethiopia in the past.
Research on the long-run effects of shocks and crises on the health, socioeconomic outcomes and human capital of individuals has developed rapidly in historical demography and economic and social history in recent decades. A key factor has been the acceleration of technologies available for generating and matching data from various historical sources. A fine-grained research corpus has elaborated the negative impact of different kinds of shocks based on the type of event, the life stage of the affected individuals, and the outcomes looked at. While so-called technophysio evolution theory traditionally emphasized the role of nutrition and economic factors in damaging human capital, recent empirical evidence from historical demography suggests that health shocks exert the most significant influence on long-term outcomes. Even when the primary shock is non-epidemiological (e.g., a crisis in livelihoods, war or incarceration), health is typically the primary dimension in which such scars are formed and perpetuated throughout the life course. Such findings could carry high policy relevance.
A classic issue with correctly measuring such effects is the interplay of scarring (long-run damage) and selection (the immediate elimination of more frail individuals from the data by the crisis biasing results upwards). In real-world contexts, causal factors typically cumulate and overlap, and so do their impacts along several dimensions. The interaction of different factors, including different types of crises, gender and inequalities in socioeconomic status (SES), has been identified as a research area where more work is needed. In our ongoing Academy project A Scarred People, building on individual level data construction from historical urban settings in 20th century Tampere and Helsinki, we are focusing particularly on the impact and interaction of three shocks: the 1916 typhoid epidemic; the 1918 Civil War; and the employment crisis of the Great Depression of the 1930s. We are able to look at the heterogeneity in impact of each of these by conditioning on SES (occupation, residence), sex, and age at occurrence. On the outcome side, in addition to end point variables like death, we can look at entire trajectories and life courses over time.
In this talk, I will provide empirical examples of the early challenges of identifying the longevity effects of two factors, being exposed to a Typhoid epidemic in 1916 and being a member of the Red Guard in 1918 during the Civil War.
Sakari Saaritsa is a Professor of Social History at the University of Helsinki. His research interests include the quantitative history of human development (particularly health, education and physiological capital), social inequality, historical indicators of well-being, and relationships between economic and human development over time. He is working with several historical datasets on Finland with local population and individual level data on demographics, anthropometrics, health and education, and involved in efforts to build national historical data infrastructure with Nordic partners. His research has been published in journals including the European Review of Economic History, Social Science History, The History of the Family and Cliometrica.
Conversation Analysis (CA) is an empirical method for investigating social interaction as it unfolds on a turn-by-turn basis. Although the method has its roots in sociology, it has also significantly influenced other disciplines, including linguistics. As an empirical method focusing on the details of interaction, CA imposes specific requirements on the data used: the data should be recorded in authentic interactional situations. Earlier CA studies have typically used audio recordings, such as (landline) telephone conversations; this has generated considerable knowledge about how various linguistic resources and elements contribute to meaning-making in social actions. Recently, however, the focus has expanded from linguistic to other resources, leading to what has been termed an ‘embodied turn’ in CA research. The increasing interest in the embodied and material resources utilized by participants necessitates new data requirements. In this talk, I will briefly present CA’s analytic framework, grounded in the details of interaction, and introduce a new dataset collected for the purposes of multimodal interactional analysis.
Salla Kurhila is a professor in Interactional Linguistics at the University of Helsinki. Her research interests cover both basic and more applied research, ranging from exploring various phenomena in social interaction to developing interventions to enhance language learning.
Computing is central to academic research, but the growing complexity of software development, the scale of current data sets, and the introduction of AI methods can be overwhelming for many research groups. To address these challenges, a new Research Software Engineering (RSE) group was recently established at University of Helsinki to provide hands-on specialist support for software development, high-performance computing, and data management. We help researchers make software more efficient and reusable, improve reproducibility, and gain confidence with computational methods. In this talk, I will introduce the RSE model, explain how it differs from traditional IT support, and describe the services we offer, including online help sessions, grant consultations and software development projects. RSE at University of Helsinki is still evolving: we invite researchers to help shape our services and tell us what kinds of computing support would benefit their work.
Alan Medlar leads the Research Software Engineering group at University of Helsinki. His research focuses on understanding and improving how people interact with information systems, such as search engines, recommender systems and AI-based applications. Recent work includes critical studies of evaluation practices in the recommender systems research community, and the development of novel interaction techniques to support navigation and increase user engagement in virtual reality.
A number of recent papers have used unsupervised image clustering methods for social science questions. However, there has not yet been research examining whether the clusters resemble the kinds of groupings that social scientists identify when conducting traditional image analysis. In response, we compare the clusters emerging through previously recommended unsupervised image clustering methods in social sciences with the manually categorised groupings to determine which, if any, of these methods are able to imitate the ‘social scientist’s gaze’. We use four different conceptual lenses from previous work, theme, content, purpose and type and use various metrics to assess the similarities. We show that none of the methods can consistently match manual analysis. Therefore, while unsupervised methods for image clustering can identify visually similar images, this may not be sufficient to capture the full breadth of methodological and theoretical approaches common for social sciences. Our work suggests that social scientists should consider unsupervised approaches as a distinct gaze or focus on developing methods which align with previous social scientists’ gazes.
Adeline Clarke is a doctoral researcher in the Social Computing Group at the University of Helsinki where she is researching multimodal large language models for use in qualitative content analysis in the social sciences. Prior to living in Finland, Adeline worked as a data scientist within industry and the Australian public service.
Communication tools enhanced with Large Language Models (LLM) are important for facilitating effective group conversations in digital workspaces, and it's crucial to develop these models to facilitate many-to-many conversations as well. Recent communicative AI applications are designed for one-to-one interactions in the form of chat. Our research investigates the challenges and opportunities of fine-tuning and deploying an AI-driven bot, Effervesce, within a multi-member group chat environment. We tested various open-source LLMs, which showed limited capabilities in handling complex, multi-actor conversations. In the current prototype, we have employed the open-source Mistral 7B model, fine-tuned using the QLoRA framework, and a dataset of 1.6k Slack messages extracted from a group conversation. To test the bot, we have conducted experimental workshops with around 40 participants in total. According to our quantitative and qualitative evaluations, the fine-tuned model results in an improved performance in following conversation structure and engagement in group discussions. Yet, the workshop participants were quite critical concerning the communication capabilities of the bot, including aspects like turn-taking and communication style. Currently, we are building an AI agent-based system that is able to better adapt to a human communication environment based on the feedback received from workshops.
Salla-Maaria Laaksonen (D.Soc. Sc) works as a Senior Researcher at the Centre for Consumer Society Research at the University of Helsinki, and she is a Docent in Media and Communication. Her research concerns communication and organizing in the platform society; business-society relations in the hybrid media system, digital organizing, and the use of data and automation in organizations.
Erjon Skenderi (PhD) is a postdoctoral researcher at the University of Helsinki. His research is focused on text representation methods, recommender systems, NLP and data analytics, with a special focus on applications of Large Language Models and AI agents in organizational settings.
Please note! This is a special event in place of the regular Brown Bag Seminar. Sign up is required and the location and time are different. Details can be found below.
You are warmly invited to the kick-off event of the AI for SSH Research network on 29 January 2026. The initiative is led by the AI researchers and leadership of the HSSH institute and the DIVSOL profiling area. Its aim is to establish a collaborative network to advance the use of AI for research in social sciences and humanities and to foster the exchange of ideas and experiences on the topic.
The meeting will take place on Thursday, 29 January 2026, from 13:00 to 17:00 in Kaisa Hall (Kaisa-sali), located on the seventh floor of the Kaisa Library (Fabianinkatu 30).
Talks will be given by Arto Klami, Sofoklis Kakouros, Matti Pohjonen, Eetu Mäkelä, and Anton Berg. The programme will also include workshop sessions dedicated to planning the network’s activities, as well as a coffee break for informal discussions.
Methodology oriented social scientists are increasingly using language models to examine the society. As large language models embody societal values and perspectives, practitioners and scholars might not account for the biases that these models cause in use. Instead of seeing biases as a source of error with big data, we could think of them as theoretical perspectives in the sense of a world-view (Weltanschauung), such as rational choice theory, Marxist theory, and feminist theory. To highlight potential biases in large language models and their implications for society –- and explore the opportunity to use language models for social science theory -- we fine-tune large language models with a specific Weltanschauung. Specifically, we incorporate the writings of Marx and Engels to fine-tune these models, aiming to infuse them with a Marxist vocabulary and world-view. We show that these fine-tuned models depict a society where affluence loses its importance, but at the same time the models are more sensitive to capitalism and economy than a standard baseline model. This investigation underscores the need to examine the values embedded in large language models before they are used to analyse society, and highlights the opportunity to incorporate a particular theoretic stance into them.
Matti Nelimarkka leads the Helsinki Social Computing Group, a collective that works in social and societal computing. His work is focused on politics, democracy and human-computer interaction, including topics such as social media politics, design as political action and societal impact of algorithmic systems. His expertise in social data science and computational methods bridges the gap between advanced methods and and social science research, with a particular focus on the interplay between social theories and their practical application through coding workflows as well as questions on validity and reliability with these novel methods. His multidisciplinary endeavors are grounded in his affiliations: he is an university lecturer at the Faculty of Social Sciences, University of Helsinki and visiting scholar at the Department of Computer Science, Aalto University. Dr. Nelimarkka holds a PhD in computer science (University of Helsinki) and the
Podcasts have gained global popularity with steadily growing number of podcast channels and recent survey studies reporting increasing levels of podcast listening especially among younger audiences (Reuters Institute Digital News Report 2025; Radio Media 2024). The rise of podcasting has also attracted attention within media scholarship which has been mainly focusing on content production and the role of podcasts as part of wider media ecology. Despite this increased scholarly interest, little is known about the motivations of podcast listeners to seek information about current affairs through podcasts. This gap in research is reinforced by major audio and podcasting platforms, such as Spotify, Apple Podcasts, and PodMe, that provide only limited data on podcast listening and the audiences that listens to podcasts. Data donation and data mirroring have emerged as methods to study media consumption across the digital public sphere, including podcasts (Jurg et al. 2024; Sejersen & Lai, 2025). In data donation—a method based on the recruitment and voluntary participation of research subjects—the researchers gain access to individual media users’ Data Download Packages (DDPs) provided by the platforms under GDPR requirements. The data is then processed and analysed for data mirroring, where the data is used to activate subsequent research methods and activities, such as in-depth or focus group interviews with the research subjects, in order to gain a deeper understanding of the motivations behind the specific media use (Jurg et al. 2024). This Brown Bag seminar sheds light on the current ethical and methodological challenges in the study of podcast audiences and podcast listening and offers an opportunity to share ideas and experiences on data donation and data mirroring as methods for studying digital media consumption in the era of closed APIs and limited data access.
Viljami Vaarala is a doctoral researcher at the Swedish School of Social Science, University of Helsinki. His PhD examines the construction of journalism’s epistemic authority by focusing on the debates and controversies surrounding the concept of journalistic truth. As part of the PhD project, he has conducted a case study on independent podcasting content on YouTube in which the concept of truth has been invoked to challenge legacy media and to establish epistemic authority for podcasting. He is part of a research team developing a project to study the role of podcasts in informing citizens about current affairs and creating senses of belonging to (political) communities among listeners.
Joint action coordination relies on the dynamic alignment of cognitive and neural processes between individuals. Recent research in social neuroscience suggests that interbrain synchrony—temporal alignment of neural oscillations across brains—supports such coordination by enhancing shared attention, prediction, and motor coupling.
Joint action coordination is a part of research agenda of The HiPerCog group led by Prof. Ben Cowley. To move beyond correlational evidence, one of the HiPerCog studies introduces a dual-transcranial alternating current stimulation (dual-tACS) protocol designed to causally modulate interbrain synchrony during joint tasks. By inducing phase-aligned oscillations in the alpha and beta frequency bands between two participants, the approach aims to test how externally driven synchrony influences behavioral coordination in real time.
The experimental design combines two paradigms: a rhythmic finger-tapping task that probes sensorimotor coupling, and a joint steering task that requires continuous co-regulation of movement. Together, these methods enable systematic investigation of how frequency-specific interbrain phase alignment contributes to the neural basis of coordinated action.
Ylätalo, H., Rodionov, A., Kilpeläinen, S., Möttönen, R., & Cowley, B. U. (2025, January 6). A Dual-tACS Study Protocol to Investigate Online Effects of Induced Interbrain Alpha and Beta Synchrony on Joint Action Coordination.
Benjamin Ultan Cowley is Professor of Learning in Humans and Machines at the Faculty of Educational Sciences, and a Docent of cognitive science. He defended his PhD in Computer Science at the University of Ulster, Northern Ireland, in 2009, and has led his group HiPerCog since 2019. The HiPerCog group studies how people learn to perform highly-demanding dynamic cognitive tasks, using computational cognitive neuroscience methods.
Andrei Rodionov is a brain researcher currently working at the intersection of neuroscience, cognitive science, and learning. He obtained his PhD in Medicine from the University of Helsinki in 2020 and he completed his postdoctoral research at Beth Israel Deaconess Medical Center Harvard Medical School in 2023. In his studies, he uses neurocognitive, electrophysiological, neuroimaging and brain stimulation methods. He has contributed to a range of clinical and basic research fields including spinal cord injury, delirium, neuroplasticity, emotion, attention, and creativity.
Hanna Ylätalo is doctoral researcher whose project aims to elucidate the role of neural synchrony in collaborative learning through causal experimental manipulation using transcranial alternating current stimulation. She has a Master’s degree in cognitive science, with a thesis on Empathy and EDA synchrony in VR collaboration, and a rich study history from diverse fields, e.g., statistics, molecular biology, experimental psychology, human-computer interaction and group processes. Hanna believes think that science at its best adopts theories from related fields to provide reliable, holistic descriptions of the world.
Science stereotypes rely on presumptions of anthropocentrism and certainty. The AI tools in research are not just a black-box, semi-secretly used. Generative artificial intelligence can also have positive impact on research. Their intelligent use is not about saving time but getting better research done. In this presentation experts discuss how to build and adapt a Large Language Model (LLM) pipeline for research work. LLMs can be turned into a machine that not only annotates in a human way but also adapts theoretical literature to provide readings of the data for research use: an example here is the Populism Theory by Ernesto Laclau. Presenters account for how they have developed and are putting into use a multi-modal LLMs pipeline to serve several research questions in multiple research projects: primarily the Horizon Europe projects on social contracts and grievance politics. They demonstrate how the instrumental use of LLMs can have value in research life. How to do it - and what does it all mean?
A further question to ask is: how does the AI use change practices in the social-sciences and humanities (SSH) field? When discussing the university's role in speer-heading the development of LLM-based research the presenters acknowledge the expertise that is needed, the necessity to build on and contribute to open-source software development, and the infrastructure of running processes locally. Need for research labs and cumulative expertise would be important to acknowledge when moving deeper into the new era that has a lot to be concerned about, much to offer - but also requires new skills and tools.
Tomi Toivio is working in the HSSH as data analyst, developing a LMM pipeline and ways to scrape data in the post-API research context for the CO3 (
Emilia Palonen is Datafication Research Programme Director at the HSSH and Associate Professor in Political Science and the leader of the Helsinki Hub on Emotions Populism and Polarisation research groupi. The project's data is drawn from the European Parliamentary elections in 2024 from Instagram and TikTok using an post-API data-collecting strategy that also has been uncovering the feed of 30 synthetic and 30 organic profiles that drew data for ten European Union member states for four weeks.
Climate disinformation is often assumed to circulate within ideological echo chambers. But what if it doesn't stay contained? This talk presents findings from analysing over 12 million tweets during COP26 and COP27, revealing how climate disinformation actively flows outward across community boundaries through unexpected pathways and unlikely intermediaries.
Using a trained machine learning classifier on validated climate datasets, we assigned disinformation probabilities to tweets denying or undermining scientific consensus on climate change. Community detection algorithms identified distinct user clusters, and social network analysis traced information flows between disinformation, mixed, and non-disinformation communities.
Most striking is the discovery of inadvertent brokers; users becoming disinformation vectors without seemingly intending to. A journalist sharing a video of a tired-looking president at COP26 became a top disinformation mediator as climate skeptics weaponized the clip for conspiratorial narratives.
The talk addresses: To what extent is disinformation spread driven by content virality versus network structure? How do we account for users amplifying disinformation without subscribing to it? And what does it mean for platform governance and policy interventions when the most influential disinformation vectors might not be the ideologues but the unintentional amplifiers? These findings challenge prevailing assumptions that toxic content is siloed, revealing instead a diffuse and dynamic information ecosystem. Understanding these brokered pathways shifts the focus from simply removing bad actors to anticipating how mainstream engagement can legitimise problematic narratives.
Simon Lindgren is a Professor of Sociology at Umeå University in Sweden, where he is also the director of DIGSUM, an interdisciplinary research centre studying the social dimensions of digital technology, and the editor-in-chief of the Journal of Digital Social Research. His research is about politics, power, and resistance at the intersection of society and digital technologies. He uses critical discourse approaches, computational text analysis, and social network analysis to study issues relating to movements, mobilization, opinions, and identities.
Antimicrobial resistance (AMR) is typically framed as a biomedical and microbiological problem, yet the conditions under which resistance emerges and circulates are profoundly social and infrastructural. The Finnish Centre of Excellence in Multidisciplinary AMR Research (FIMAR) is a multidisciplinary consortium comprising five disciplines: clinical microbiology, environmental microbiology, evolutionary biology, bioinformatics - and ourselves, Sociology. Together, we seek a paradigm change in AMR research by developing a method to integrate microbiological and sociological data for the purposes of understanding AMR spread and evolution across various scales from the molecular to the level of political economies. Often, qualitative and quantitative methods are siloed, both in terms of practices within academia and in terms of journal publications. Problems like AMR which are micro-bio-social and therefore multiple by nature necessitate moving beyond these silos. As the ultimate aim of FIMAR is to integrate AMR data from multiple disciplines, we have taken a mixed-methods approach to our data set that draws on both qualitative and quantitative analyses. At stake in this work, situated as it is in a multidisciplinary context, is how to hold complexity across multiple epistemological registers in such a way that enhances our shared understanding of AMR.
As such, we developed a multi-model study to recruit 'index' households (diverse by setting and socioeconomic risk) in Cotonou, Abomey-Calavi, and Grand-Popo (Benin) to characterise how exposures arise in everyday life and how these relate to microbial findings. As the sociological component of the research project, we conducted in-depth interviews and mapping exercises with 9 index households on daily routines, practices and understandings of microbial riskiness. Our analysis thus far have led us to conceptualise 12 dimensions of practices participants discussed in relation to microbes and their daily lives and we are continuing to explore analytical avenues to push our findings further. We hope that this work-in-progress talk will invite discussion, comments and questions about using mixed methods approaches in hopes that this will support and invigorate future efforts in this area.
Iona Walker is a medical anthropologist fascinated by the poetry and science of being human in a more-than-human world. Iona completed her PhD in social anthropology at the University of Edinburgh (funded by the Wellcome Trust) where she conducted ethnographic fieldwork with scientists at British research-intensive university during the Covid-19 pandemic. This work explored how scientists imagine, respond to and understand human-microbe relationships in the context of their research on antimicrobial resistance and respiratory tract infections. In particular, Iona’s research explores how researchers reconfigure their understandings of health, infection and human-microbe relationships, through their AMR research, away from antagonism and toward more situated, emergent forms of relations; as well as how scientists construct and articulate their everyday research practices as ‘AMR’ through the imperatives of academic knowledge production in the UK. Iona is a founding member of Beyond Resistance, an interdisciplinary network designed to catalyse curiosity and collaboration between the sciences, arts and humanities in regards to microbial challenges like AMR.
Artificial intelligence and big data systems have been widely criticized for the actual and potential harms they produce. In response, policymakers have introduced legislation to mitigate these risks. Industry narratives of responsible or trustworthy AI often aim to reassure citizens and regulators, while recent U.S. policy developments have instead dismissed ethical guidelines, fairness tests, and legal constraints as innovation-stifling burdens. In contrast, despite the European Commission’s recent retreat from strong AI regulation, public support for robust safeguards remains high across the EU.
Beyond landmark legislation such as the AI Act, a complex social web of actors, institutions, and practices has emerged to contest and shape algorithmic systems. This talk focuses on the Netherlands as a case study to examine how custodianship for algorithmic systems is being developed within public management. It explores the rise of formal education in AI literacies, the professionalization of roles dedicated to data and AI ethics, and the growing importance of government-mandated supervisory authorities. These forms of custodianship intersect with broader practices of contestation—spanning activism, advocacy, investigative journalism, and political deliberation, often at the municipal level.
Finally, the talk reflects on the role of action researchers in critical data and AI studies, examining how they can contribute to and strengthen these emerging practices. Drawing on the recent publication Collaborative Research in the Datafied Society: Methods and Practices of Investigation and Intervention (Schäfer, Van Es, Lauriault, 2024), it invites discussion on the opportunities and challenges of collaborative and action-oriented research in shaping a democratic digital society.
Discussant: Minna Ruckenstein
Mirko Tobias Schäfer is Associate Professor of AI, Data & Society and Science Lead of the Data School Utrecht University. His research explores the societal impact of datafication, algorithms, and artificial intelligence, with a particular focus on public management, and citizenship. He is co-editor of Collaborative Research in the Datafied Society. Methods and Practices of Investigation and Intervention (Amsterdam University Press, 2024). Dr. Schäfer also serves on the Advisory Board Analytics of the Netherlands' Ministry of Finance. For the past three years, he was a Visiting Professor at the University of Helsinki’s Institute for Social Sciences and Humanities.
We all want to read less so we make robots do the reading. Sadly, the lazy language models do not have the patience to read the long texts either, and so they read a bit from the beginning, a bit from the end and then make up the rest. How much time is actually saved, if the robot reader is not reliable enough and one has to fact check it and essentially read the long texts?
This talk presents an extension to the findings of the paper “
Erkki Mervaala is a researcher at the Finnish Environment Institute’s Politics of Knowledge group, and a PhD student at the University of Helsinki. His research interests include climate change & economic growth representations and intersections in the media, and the buzz and hype around large language models and generative artificial intelligence.
Whose Angels? is a project funded by the Kone Foundation that integrates photographic art with interdisciplinary research, encompassing the study of religions, art history, and folkloristics. The project aims to investigate contemporary vernacular and popular imagination and worldviews. We utilize contemporary fine-art angel-themed phogographs, created by our artist Hanne Kiiveri, as prompts to elicit thoughts and feelings from diverse audiences. These photographs have been exhibited in various venues and presented to a range of audiences during workshops. The responses—opinions, memories, critiques, and associations—gathered from viewers serve as research material, primarily documented in real-time. Additionally, some audience members have volunteered to collaborate with the artist in designing new angel images, and this creative process has also been recorded as part of our research, including interviews, observations, and the final artworks. One of the objectives of our project is to explore and assess this kind of ethnographic methodology (in which we first construct our case and then study it) and its potential for future research.
Terhi Utriainen is professor in the Study of religions at the University of Helsinki. Her interests include ethnographic study of contemporary vernacular religion, religion, gender and embodiment, and ritual studies. She has recently directed the project Learning from new religion and spirituality (funded by the Research council of Finland 2019-2023) and presently directs the 3,5-year project Whose Angels? funded by the Kone Foundation.
Oscar Ortiz-Nieminen holds a PhD in Art History and a MTh in the Study of Religion. He is currently a postdoctoral researcher at the University of Helsinki. His research interests lie in the intersections of religions and worldviews with visual culture and with the built environment.
This talk presents both empirical findings and methodological insights from the Belgian interuniversity research project TACOS, which investigates preschool teachers’ language-supporting competencies through a multimethod approach. The study uses mobile eye-tracking (MET) to examine, before and after a professional development intervention, the question: Which children are being overlooked, and with what consequences for their language development opportunities?
In a randomized controlled trial with a posttest-only comparison group, we captured 1,300 minutes of classroom interaction involving 65 preschool teachers and 575 children. Multilevel negative binomial regression models revealed systematic inequalities in teacher attention: linguistically vulnerable children—those perceived by teachers as having low speaking confidence, weaker language skills, and/or a different home language—received less attention. Importantly, results provide novel evidence that teachers’ attention allocation can be positively influenced through targeted professional development.
The talk will address the following themes:
Thibaut Duthois is a Belgian doctoral researcher at Ghent University, working on educational inequality and teacher professionalization under the supervision of Prof. Ruben Vanderlinde, Prof. Maribel Montero Perez, and Prof. Piet Van Avermaet. His main research interest lies in early childhood education, with a particular focus on uncovering mechanisms of inequality at the teacher level. To this end, he employs mobile eye-tracking technology to study teachers’ visual attention and interaction patterns in the classroom.
The discussion on ethical challenges is timely because co-research is becoming more common in various disciplines. However, there is no clear instruction on the ethical procedures, which is a concern raised by TENK, as well. This panel discussion of four experts at the University of Helsinki will commence with brief introductions from each presenter. This Brown Bag seminar is part of a Catalyst Grant project led by Auli Vähäkangas, who will also moderate the discussion.
Meri Kulmala is a sociologist and a specialist in inequality studies, with extensive experience in co-research involving various marginalized and minoritized groups. She serves as Research Director at the Faculty of Social Sciences, University of Helsinki. Among her responsibilities, she leads two interdisciplinary research networks: the Helsinki Inequality Initiative (INEQ) and Resilient and Just Systems (RESET). She also leads the SRC OBaMa Consortium, which co-investigates and co-creates solutions to inequalities in democratic participation together with marginalized and minoritized groups. She is also a founding member of the Finnish Co-research Network. Her interests in co-research in relation to research ethics focuses particularly on identifying and addressing situations where the core principles of co-research—such as participation, power-sharing, and valuing lived and embodied expertise—come into conflict with the structural constraints of academia. These tensions often emerge throughout the research process: beginning with the inclusion of lived experience in the design phase, continuing through the balancing of power during empirical work itself, and culminating in the dissemination stage, where issues of ownership and authorship become especially pronounced.
Reetta Mietola a university researcher in RESET – Resilient and Just Systems, at Faculty of Social Sciences, UH, and the deputy PI in the SRC funded OBaMa project, studying barriers of democratic participation with marginalized and minoritized groups. Her interest in co-research was initiated by disability studies, more specifically inclusive research developed in this field. She is particularly interested in how co-research challenges ideas and practices taken for granted in academic research: who can lead research, what kind of capacity is required from a researcher and in relation to this, what is needed for research to be inclusive. This also presents challenges for research ethics: how to balance protection with access to participation; what does informed consent mean in practice and in a co-research process; how are risks and related responsibilities negotiated within the research team; how to deal with authorship, and what kind of ethical consideration does this involve.
Auli Vähäkangas is Professor in Practical Theology and Vice Dean at the Faculty of Theology, University of Helsinki. She is a member of the Research Ethics Committee in the Humanities and Social and Behavioural Sciences at the University of Helsinki as well as a member in RESET and INEQ. Vähäkangas’ research has focused on death and dying and on people in vulnerable situations. Presently she is leading Meaningful Deathscapes: Worldview minority cemeteries in Finland (MeDea, 2024-2028). Ethical review practices are also challenged by collaborative research. It is important to discuss how far the current ethical guidelines fit the settings of collaborative research, and to what extent we can assume that different research collaborators adhere to our research ethics. Vähäkangas acknowledges the risks existing with the researchers’ roles and questions of power in connection to co-research method. Questions of power are multidimensional, both between academic and participant researchers and between those PRs who are in a leadership position in the worldview communities and those who have a member status in them. Questions of power bring ethical challenges to the co-research project from the planning stages to the dissemination of research, but especially challenging are the power relations during the field work.
The digital information environment is changing rapidly, with challenges such as disinformation, generative AI, and the spread of short-form video platforms. Traditional approaches to teaching information literacy often struggle to keep pace with these shifts or to reach people outside specialist or academic settings. This presentation introduces Infodemic Inc., a digital educational game developed as part of the InfoLead project, a collaboration between the University of Helsinki, the University of Oxford, and the University of Florence. Alongside executive training for judges and policymakers and a toolkit and casebook, the game is one element in the project's efforts to build professional networks and strengthen media literacy.
The presentation also asks what role games can play in communicating research results and reaching wider audiences. Can complex academic and legal debates about platform governance be made accessible through playful interaction without losing their nuance? How might game-based formats complement, or even challenge, more traditional ways of sharing research and knowledge? Drawing on the design and early testing of Infodemic Inc., the session considers both the potential and the limits of games as tools for popularizing research and supporting media literacy. This session launches the new game/simulator developed by the InfoLead project but also discussed its wider ramifications for changing academic practices.
Dr Aleksi Knuutila is a University Researcher at the Department of Sociology at the University of Helsinki and the Helsinki Institute for Social Sciences and Humanities. After his doctorate in the Digital Anthropology programme at University College London, Knuutila’s research has focused on online harms such as misinformation and harassment and how political groups take advantage of contemporary information environments. His current research projects focus on developing tools and infrastructure for journalists working on conflicts and applying generative AI to interpretative research workflows.
Dr Matti Pohjonen currently works as a Senior Researcher for the Helsinki Institute for Social Sciences and Humanities (HSSH), University of Helsinki, working with methodological development on the use of internet and social media data, including debates on generative AI and LLMs. He also currently co-leads The EU Horizon-funded project ARM, which focuses on information suppression and information freedoms in China, Russia, Ethiopia and Rwanda and InfoLead, a digital literacy project developed together with the University of Oxford and University of Florence.
The Center for Research Data and Digital Scholarship (CRDDS) at the University of Colorado Boulder is an interdisciplinary research center focusing on research data infrastructure and expertise. CRDDS is built on a partnership between the University Libraries and Research Computing, and represents an uncommon fusion of units on the academic and operational sides of the university from various disciplinary and professional backgrounds with a common mission to empower researchers navigating the research data lifecycle on campus. This work is achieved mainly through collaborative grant opportunities, workshops and seminars, and certificate and micro-credential programs. In addition to giving an overview of our work, I will outline the ways in which CRDDS is experiencing the shifting federal research and higher education landscape at the university and within the scientific context in Boulder, which includes university centers, labs, institutes, and other units as well as federal laboratories.
Thea Lindquist is Professor and Executive Director of the Center for Research Data and Digital Scholarship at the University of Colorado Boulder, an interdisciplinary center specializing in expertise and infrastructure for data-intensive research and education and in open publishing. Her research interests include integrating historical and computational approaches in the study of 17th-century European history and data curation for interdisciplinary and highly collaborative research.
Philosopher Achille Mbembe once remarked that Africal holds a paradoxical position in modern formations of knowledge. On the one hand, he writes, “it has been largely assumed that "things African” are residual entities, the study of which does not contribute anything to the knowledge of the world or of the human condition in general.” Yet, on the other hand, he also notes, Africa has always been also perceived as a kind of laboratory that can help “gauge the limits of our epistemological imagination or to pose questions about how we know what we know and what that knowledge is grounded upon (Mbembe, 2010, p. 654).”
University of the Witwatersrand (Wits) and the University of Edinburgh recently launched a new SARChi SA-UK Bilateral Chair in Digital Humanities, appointing Professor Gagliardone as its inaugural holder. The aim of the 5-year Chair is to advance more inclusive and equitable digital futures in social sciences and humanities research globally. Building on the long critical tradition of research in South Africa, the Chair aims to reposition Africa as hub of conceptual, technological and methodological innovation amidst the current boon in the use of AI in social sciences and humanities by highlighting how historically marginalised forms of knowledge can inform humane and socially just approaches to digital transformation and digital humanities research.
Matti similarly recently received Africa Programme seed funding for his ongoing project “Generative AI and Africa: New Methodological Directions for Social Sciences and Humanities Research (GAINS).” This project aims to critically explore the use of generative AI and large-language models (LLMs) for social sciences and humanities research in Africa. By experimenting with custom-trained AI models and fostering collaboration between Finnish and South African institutions, it seeks to enhance research methodologies and facilitate knowledge exchange globally.
Building on this ongoing dialogue between HSSH, University of Helsinki and University of Witswatersrand, this talk openly discusses some of the challenges in such North-South collaboration in social science and digital humanities. What are some of the theoretical, methodological and empirical questions that are raised and how can critical research potentially help address this? What are the limits of our epistemological imagination?
Iginio Gagliardone in the inaugural SA-UK Bilateral Chair in Digital Humanities at Wits University in Johannesburg, South Africa, and a fellow of Wits’ Machine Intelligence and Neural Discovery (MIND) Institute. He is the author of “The Politics of Technology in Africa” (2016) and “China, Africa, and the Future of the Internet” (2019). His most recent work examines the international politics of Artificial Intelligence and the emergence of new imageries of technological evolution in Africa.
Matti Pohjonen currently works as a Senior Researcher for the Helsinki Institute for Social Sciences and Humanities (HSSH), University of Helsinki, working with methodological development on the use of internet and social media data, including debates on generative AI and LLMs. He also currently co-leads The EU Horizon-funded project ARM, which focuses on information suppression and information freedoms in China, Russia, Ethiopia and Rwanda as well as InfoLead, which is a digital literacy project and an online tool targeting judges and policymakers developed together with the University of Oxford and University of Florence.