Research on Text Credibility Assessment Presented at the International FCTA 2026 Conference

The international FCTA 2026 conference (International Conference on Fuzzy Computation Theory and Applications), part of IJCCI (International Joint Conference on Computational Intelligence), took place in Angers, France, on 28–30 October 2026. Petra Števuliáková from the Institute for Research and Applications of Fuzzy Modeling at the University of Ostrava presented research findings developed within the project “Biography of Fake News with a Touch of AI”.

FCTA is an international scientific conference focusing on fuzzy methods, computational intelligence, and their applications. As part of IJCCI, it brings together researchers and experts working in fuzzy systems, evolutionary computation, neural networks, and artificial intelligence. Explainable and generative artificial intelligence were also among the important topics addressed during the conference.

Petra Števuliáková presented the paper “A Fuzzy Approach to Text Credibility Assessment via Semantic Attributes”, which introduces a method for evaluating the credibility of textual information. The research addresses the increasingly relevant challenge of disinformation and manipulative content, particularly in the context of rapid advances in generative artificial intelligence.

The proposed method combines large language models (LLMs) with fuzzy approaches. Language models are used to assess selected semantic attributes of a text, which are subsequently evaluated using fuzzy methods. The aim is to assess the degree of text credibility and identify characteristics that may indicate misleading or manipulative content. An important feature of the proposed approach is its interpretability, allowing researchers to explain which text attributes contributed to the final assessment.

The presented paper builds on previous research findings introduced at IPMU 2026 and ITAT 2026. The research is being conducted in collaboration between Petra Števuliáková, Jiří Janeček, Petr Sosík, and Michal Místecký, bringing together expertise from two research work packages within the project. This interdisciplinary cooperation enables the integration of knowledge from artificial intelligence, fuzzy modeling, and textual information analysis.

The presentation at FCTA 2026 represents another important step in developing methods for text credibility assessment. Future research activities include validating the proposed approach on additional datasets and integrating it with complementary credibility assessment methods.

The research is a direct outcome of the project “Biography of Fake News with a Touch of AI: Dangerous Phenomenon through the Prism of Modern Human Sciences”, registration no. CZ.02.01.01/00/23_025/0008724, and contributes to the development of tools supporting a better understanding of information credibility in the digital environment.

Updated: 05. 11. 2026