"Jag tänker att det är sant... tills någon säger att det är fejk" : En kvalitativ och explorativ studie om källkritik och bedömning av deepfakes på TikTok

Sammanfattning

Deepfakes, manipulated media created using artificial intelligence, pose a significant challenge to trust credibility in digital content. This study explores how young adults evaluate authenticity of such content and assesses the potential of technical inventions, such as Community Notes, to foster critical thinking and mitigate the spread of misinformation. Using a qualitative method, including semi-structured interviews along with controlled experiments, this paper examines user behaviours and preceptions. The findings revealed that while participants are aware of the risks associated with manipulated content, they often rely on collective verification methods, such as reading comment sections, to determine credibility. This behavior reflects a reliance on crowdsourcing, highlighting the role of social cues in addressing misinformation. However, the results also emphasize the need for accessible and user-friendly technical tools to enhance and inspire critical evaluation in relation to manipulated content. The study concludes with recommendations for future research to support users in navigating the growing prevalence of deepfakes on social media platforms.