<p>Recent literature has highlighted heightened ethical concerns about autonomy in AI-driven nudges compared to traditional nudges. This article argues that a clearer understanding of these concerns requires distinguishing between two distinct aspects of autonomy. The first relates to recognizing a reason as the cause of a preference (self-agency-based autonomy). The second pertains to the alignment between preferences and actions (autocracy-based autonomy). This distinction clarifies that the heightened concerns about autocracy stem from reasonable conjectures rather than established threats.Furthermore, this article highlights how AI systems, beyond serving as persuasive tools, have the potential to protect individuals from undue persuasion. Specifically, AI can help categorize and identify choice environments that undermine individual autocracy. The key advantage of AI in this context is its ability to surpass intuitive methods in evaluating the impact of nudges on autonomy. Ultimately, this article contributes to a more nuanced ethical framework for assessing AI-driven nudges. It challenges some of the overly negative narratives surrounding AI in this domain, offering a more balanced and constructive view.</p>