Title: The Deception of Choice in AI-Driven Critical Decisions

Abstract:

Leveraging Ullmann-Margalit’s framework of "opting"—defined by its transformative nature, irrevocability, and the presence of foreclosed alternatives—this study reveals a significant ethical oversight in current AI ethics discourse. We identify a phenomenon we term the "illusion of opting," wherein individuals and communities face a misleading sense of meaningful agency in consequential matters, even as their actual capacity for genuine choice is eroded.

Rather than viewing AI merely as a tool for optimizing pre-established goals, we contend that its evaluation must center on its ability to safeguard and nurture "meta-capacity." This refers to the socially and institutionally supported ability to form, challenge, revise, and claim ownership over both the means and ends of action. This perspective is particularly critical for marginalized groups, who bear the heaviest burden when AI-guided pathways divert behavior and action, leaving them least equipped to manage the consequences of the opting illusion.

To address these challenges in AI-mediated consequential decisions, we outline three normative imperatives: 1. Existential Honesty: Recognizing the inherent limitations of predictive models. 2. Ecological Rationality: Embedding guidance within the diverse contexts of lived experience. 3. Counterfactual Reparation: Identifying and remedying the loss of foreclosed alternatives when AI-driven decision processes go awry.

Source: arXiv Generated at: 2026-06-04 00:00:00 UTC