Data from 381 respondents for cognitive-behavioral modelling of the perceived plausibility of artificially generated content in Ukraine

The dataset contains anonymized results of a questionnaire survey of 381 respondents, conducted in Ukraine in April 2026, and was formed to study the cognitive-behavioral features of perceived plausibility of artificially generated content. The data are intended for testing an information technology for cognitive-behavioral modelling, which combines the formalization of linguistic assessments, fuzzy logic, neuro-fuzzy modelling, and weighted aggregation of criteria. Each record in the dataset corresponds to one respondent and includes socio-demographic characteristics, namely gender, year of birth, socio-economic status, level of education, average net monthly income, marital status, and region of permanent residence. A separate block of variables describes the respondents’ culture of information consumption: frequency of Internet use in leisure time, frequency of Internet use for work, frequency of television viewing, frequency of reading printed newspapers or magazines, types of websites visited, types of television programmes watched, and types of magazines read. In addition, the dataset contains respondents’ answers regarding their perception of several pairs of news texts related to socially significant topics, including fuel, elections, and the protection of personal data in digital services. For each topic, respondents were presented with two text versions, one of which was presented as human-created and the other as artificially generated content. The responses make it possible to assess citizens’ ability to distinguish between plausible, fake, and artificially generated news content. The dataset can be used for the formalization of respondents’ linguistic statements, the construction of normalized quantitative assessments, the training of neuro-fuzzy models, the calculation of an integral index of perceived news plausibility, and the determination of the corresponding linguistic level. The practical significance of the data lies in the possibility of identifying groups of citizens with different levels of information vulnerability to fake news, disinformation, and artificially generated content.