<p>Real persona social AI (RPSAI) systems, conversational AI trained to emulate specific individuals using their personal data, allow anyone with data access to create interactive likeness simulations. While most existing social AI literature focuses on user wellbeing, this paper examines RPSAI’s effects on the modeled individuals whose likeness is appropriated to create these systems. I argue that RPSAI harm modeled individuals through three interconnected mechanisms. First, personal data transforms from informational input into generative likeness that operates beyond traditional privacy protections. Second, RPSAI systems sever individuals from control over their social authorship by undermining selective disclosure, enabling misrepresentation, and structural reduction. Third, RPSAI systematically displaces the modeled individual’s functional value in social contexts, reducing authentic human presence to raw material for synthetic optimization. The paper proposes initial mitigation strategies including banning unauthorized RPSAI creation and implementing comprehensive disclosure requirements. As digital duplication becomes technically accessible and our social lives increasingly datafied, recognizing these compound harms across data, authorship, and extrinsic value provides essential groundwork for preserving authentic human agency against its own synthetic appropriation.</p>