Based on privacy calculus theory, we derive hypotheses on the role of perceived usefulness and privacy risks of artificial intelligence (AI) in public services. In a representative vignette experiment (n = 1,048), we asked citizens whether they would …