Abstract
As artificial intelligence (AI) evolves, conspiracy theories have emerged that authorities will use AI to oppress humanity, or AI itself will. We propose that perceived high autonomy and low interdependence of AI increase AI-related conspiracy beliefs. Four studies (total N = 1897) have examined this line of reasoning. Study 1 (N = 300) supported the hypotheses in a correlational survey. Studies 2 (N = 400) and 3 (pre-registered; N = 400) manipulated the autonomy and interdependence of AI in experiments. Both studies found that higher autonomy and lower interdependence increased AI-related conspiracy beliefs, while perceived threat to society mediated these effects in most cases. Study 4 (pre-registered) replicated findings from Study 2 in the United States (N = 400) and China (N = 397) and found cultural differences in AI-related conspiracy beliefs. These findings illuminate how the perceived properties of AI contribute to AI-related conspiracy beliefs.