Determinants of Intention to Share AI-Generated Political Dis(mis)information: An Extended Theory of Planned Behavior

Sammanfattning

Rather than viewing the spread of disinformation as a technological problem, this study models disinformation as a product of interconnected psychological and cognitive factors. These factors are derived from Theory of Planned Behavior while extending the theory with additional constructs to robustly test intention of sharing AI generated disinformation. The study leveraged quantitative data from a purposive sample of 200 participants drawn from Swedish context. The primary hypothesis was tested in a regression model, with results indicating that habit was the most significant determinant of intention to share AI generated disinformation. Beyond habit, a positive attitude toward the content and perceived veracity also significantly increased sharing intention. Conversely, perceived behavioral control exerted a significant negative influence on sharing intention revealing a compelling paradox. Subjective norms, disinformation vigilance, and political right orientation had no significant impact. Further exploratory analysis on attitude revealed that predictive power of attitude is driven primarily by its experiential value rather than instrumental utility. Additional exploratory analysis with population characteristics suggests universal vulnerability to disinformation irrespective of one's gender, age and political alignment. Understanding disinformation in AI and Swedish context could help to develop strategic communication initiatives for building psychological defense.