CEU Electronic Theses and Dissertations, 2023
| Author | Fritzen, Natalia Martins |
|---|---|
| Title | AI-Nudging and Individual Autonomy: Moral Permissibility and Policy Recommendations |
| Summary | As the deployment of Artificial Intelligence (AI) technologies increasingly triggers ethical, legal, and socio-political debates, important advancements have been made in the study of AI’s effects on decisions concerning humans and their well-being. Yet, the same is not true about AI’s effects on human’s decisions themselves. In this dissertation I address this gap by discussing how AI affects human decision-making through nudging practices. My first argument is that, from a moral point of view, AI-nudging is more concerning than traditional nudging insofar as it is more likely to impair autonomy, for it exacerbates the two conditions I take for a nudge to be morally wrong: manipulation and infantilization of the nudgee. I explain which are the features of AI-nudging that makes it morally concerning and, based on them, I make my second argument, namely that the state has the duty to protect individual autonomy, and that this can be done through two main policies. First, the state ought to portray the harm to autonomy as a harm in itself, and not (only) conditioned on material harms that the autonomy impairment might cause. Second, the state ought to regulate AI as to protect autonomy on the grounds of protection of autonomy itself, and not on other (already existing) rights, such as freedom of thought or privacy. I conclude that only in this way the abstract moral concerns around AI-nudging can be translated into feasible and effective policies to protect autonomy. |
| Supervisor | Gheaus, Anca; Soave, Tommaso |
| Department | Political Science MA |
| Full text | https://www.etd.ceu.edu/2023/martins_natalia.pdf |
Visit the CEU Library.
© 2007-2021, Central European University