Who and How Spreads Disinformation on Social Media? Researchers at the School of Digital Technologies at Tallinn University Developed a System to Detect It.

Professor David Lamas, Associate Professor Sónia Sousa and Junior Researchers Debora Conceicao Firmino De Souza, and Gabriela Beltrão from the School of Digital Technologies (DTI) at Tallinn University participated in the international research project MARTINI. The broader aim of the project was to develop artificial intelligence tools capable of automatically detecting and profiling disinformation spreaders across different social media platforms. The DTI researchers developed a taxonomic classification system that provides the foundation for building such a tool.

The DTI researchers coordinated a collaborative investigation into disinformation and misleading content circulating on online social networks, focusing on the nature and agency of those spreading false information. Researcher Debora Conceicao Firmino De Souza stated that a major challenge in this kind of work is identifying these agents, as the content circulating online is not always malicious. “A significant amount of false content is disseminated unintentionally by unaware users. The problem is further complicated by the fact that disinformation is spread both by humans and automated accounts, such as bots, and that both individuals and organized groups may be involved. Given the rapid spread of misleading content, there is often very little time available for detection and fact-checking,” Souza added.

The contribution of the DTI lies in developing a classification that enables the creation of more general and event-independent detection mechanisms capable of identifying and profiling disinformation spreaders. The taxonomical system created by the researchers consists of three parts, designed to answer the following questions:

1. Who is the disinformation spreader, and what are their characteristics? For example, do they create false content themselves or merely share it? Are they an individual or an organization?
2. What do they do and how do they operate? Do they manipulate news feeds, shift the focus of public discussions, or engage in reputation management?
3. What is their broader strategy and objective? Do they indiscriminately spam content, artificially amplify their influence, or create echo chambers?

The developed taxonomy is universal and not tied to any specific topic, country, political orientation, or event. It can be applied across a wide range of contexts to detect and profile disinformation spreaders.

The system was tested using real-life examples. Researchers analyzed Telegram channels where misleading or malicious content had been shared. The case study revealed that, in practice, classifying content is highly complex, as interpretation depends heavily on context and network. 

Furthermore, the assessment of whether something constitutes disinformation often relies on broader factors: the same post may appear neutral in one discussion but take on misleading or even harmful meaning in another. It is also difficult to determine who originally created a piece of content and who merely redistributed it, as material moves rapidly across social media platforms and reposted websites, often without attribution to its original source.

Associate Professor Sónia Sousa is confident that the work has significant value nonetheless: “The developed framework provides a reliable foundation for researchers and technology developers to continue working on artificial intelligence tools capable of detecting and profiling disinformation spreaders and to aid in the responsible design of network systems.”

Currently the project team is working on a forthcoming publication with the consortium partners to present the concluding findings of the project.

The research findings were published in an article “Naming Names: A Taxonomy of Malicious Actors and Their Activities in Online Social Networks“ in the Proceedings of ECCE 2025 and are available in the ACM Digital Library. 

This article was originally published by the School of Digital Technologies, Tallinn University.

If decoding who spreads disinformation online has sharpened your eye for the hidden mechanics of social media, take the next scroll – where influence turns from distortion to empowerment, and the same platforms begin to shape how we see ourselves. Read how Social Media Influencers Can Help Rebuild Women’s Body Confidence!