The generation of images and text by artificial intelligence (AI) is rapidly transforming digital interactions, yet research on their psychological and societal impact lags behind. Our pioneering work reveals that people frequently mistake AI-generated faces for real ones—and are not aware they are being deceived. This human vulnerability has profound implications for digital security and misinformation, with synthetic identity fraud projected to have an annual global cost of $40 billion USD by 2027. This project unites cognitive and social psychology with computer science to develop robust AI detection strategies and redefine human face perception theory. Using behavioural experiments and deep neural networks, we will model how people mentally represent real and AI-generated faces and create the first psychometric tests to measure AI-detection ability. By integrating human perceptual strengths with machine learning, we aim to optimise AI-detection accuracy to safeguard digital identities, combat misinformation, and protect individuals from exploitation. Understanding how AI disrupts social cognition will not only mitigate its risks but should also unlock new insights into human perception. This research hopes to provide policymakers, educators, security agencies, and forensic professionals aims science-backed strategies to counter AI-driven fraud and misinformation, ensuring a safer and more resilient society in the age of artificial identities.