Artificial intelligence tools for fact-checking and detecting disinformation were at the centre of the international expert seminar From Analysis to Action, organised by GATE Institute. Researchers and experts from European institutions, media organisations and research bodies discussed vulnerabilities in Bulgaria’s information environment and approaches to countering manipulative messaging through technology, media literacy and regulation.
The forum was opened by Prof. Sylvia Ilieva, Director of GATE Institute. In his keynote lecture, With a Little Help from NLP: My Roadmap to Social Good, Prof. Ruslan Mitkov from the University of Alicante and Lancaster University presented applications of natural language processing (NLP) in education, translation and support for people with communication difficulties. He noted that automatically generating test questions can be almost four times faster than creating them manually, without compromising quality.
During the session From Detection to Deployment, Prof. Preslav Nakov from Mohamed bin Zayed University of Artificial Intelligence and the Qatar Computing Research Institute explored how large language models can be used for automated fact-checking and media content analysis.
GATE experts also shared their practical experience in the field. Stanislav Penkov presented a tool for detecting logical fallacies, while Dobromir Tsolov and Georgi Goranov demonstrated an application for analysing political narratives. Participants emphasised that AI-generated results must be verified and supported by reliable evidence.
Vulnerabilities in Bulgaria’s information environment were the focus of the discussion The Vulnerability Audit. Participants included Dilyana Kirkovska from the Council for Electronic Media, Rositsa Dzhekova from the Center for the Study of Democracy, Nikola Miladinov from the European Commission Representation in Bulgaria, Iglika Ivanova from GATE and the Coalition for Media Literacy, and Mihail Mishev from GATE. The discussion was moderated by Ruslana Margova from GATE.
Experts discussed declining trust in traditional media, citizens’ disengagement from public debate and polarisation in online spaces. Topics included European media literacy policies and the protection of journalists against strategic lawsuits against public participation (SLAPPs).
The third session, The Manipulation Toolkit, focused on linguistic techniques and information influence strategies, including the use of historical arguments, emotional appeals and linguistic markers of manipulation in online content.
The discussion was moderated by Keith Kiely from GATE and featured Prof. Petya Osenova from Sofia University and the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences, Devora Kotseva and Todor Kiryakov from Identrics, Rositsa Dzhekova from the Center for the Study of Democracy, and Yana Naydenova and Nevena Grigorova from GATE.
Participants explored ways to automatically detect offensive language using linguistic resources such as HurtLex. Experts stressed that understanding information manipulation requires analysing entire narratives rather than isolated claims.
The fourth session, Inoculate Before Infection, moderated by Iglika Ivanova, focused on prebunking—an approach that prepares people to recognise disinformation before they are misled by it. Discussions covered fact-checking training, educational initiatives for different age groups and the exchange of knowledge between generations.
The concluding discussion examined European regulations, including the Digital Services Act (DSA) and the AI Act. Participants included Ivo Emanuilov and Mihail Mishev from GATE and Rositsa Dzhekova from the Center for the Study of Democracy, with Keith Kiely as moderator. Key topics included online platforms’ accountability, the protection of freedom of expression and the need for closer cooperation among institutions, technology companies, research organisations and civil society.