Introduction: The rapid proliferation of AI-generated and algorithmically curated content has intensified societal polarization, misinformation exposure, and echo-chamber effects. Psychology has a vital role in addressing these challenges by developing interventions that strengthen critical decision-making and trust in information systems.


Purpose: This study evaluated whether an AI-driven media literacy intervention could reduce susceptibility to misinformation, mitigate polarization, and influence digital behavior patterns.


Method: In a two-week randomized controlled trial (N = 40), participants were assigned to one of two groups. Both groups received a curated daily standard news feed (balanced across outlets), while the intervention group additionally received intermittent short literacy videos and micro-quizzes explaining misinformation cues, bias recognition, and algorithmic personalization. Data were collected through active surveys and passive digital phenotyping via a mobile application. Daily ecological momentary assessments (EMAs) captured trust, perceived bias, and sharing intentions. Digital phenotyping measures included app usage diversity, time spent on news content, and engagement with push notifications. Pre- and post-intervention tasks assessed misinformation detection, confidence, and affective polarization.


Results: Preliminary findings suggest that the intervention group demonstrated improved accuracy in misinformation detection, reduced affective polarization, and more diverse digital engagement patterns compared to controls. App usage diversity and reduced reliance on partisan sources emerged as potential mediators of intervention effects.


Conclusions: AI-enhanced literacy interventions show promise in fostering resilience against misinformation, reducing polarization, and encouraging responsible digital engagement. This research underscores the potential of psychology to inform digital literacy strategies aligned with democratic values and sustainable information ecosystems.