DETECTING AI-GENERATED DEEPFAKE IMAGES: STUDENTS PERCEPTIONS

The rapid advancement of generative artificial intelligence has enabled the creation of highly realistic synthetic images and videos, raising major concerns about information integrity, political manipulation, and public trust. This study investigates how effectively individuals can detect AI-generated deepfake images, focusing on human judgment as automated detection tools remain unreliable, biased, and easily bypassed. Using a mixed-methods design, political deepfake content was generated through multiple AI models and evaluated through a structured survey administered at Suleyman Demirel University. Results show that although participants frequently encounter AI-generated media, their confidence in distinguishing authentic from synthetic content does not consistently align with actual detection accuracy. Findings further reveal that deepfakes can appear highly credible and hold persuasive potential even when recognized as artificial. The study highlights the need for multi-layered responses, including improved detection technologies, stronger media literacy, and more robust regulatory frameworks to mitigate emerging societal risks.