Why AI's 'choices' might not be choices at all

When AI mediates important decisions—hiring, healthcare, loans—people feel they're choosing, but the system narrows options and obscures trade-offs, leaving them less capable of genuine agency. The authors argue AI should be judged not on optimizing outcomes but on protecting people's capacity to form and contest their own goals. For disadvantaged groups, this illusion can be particularly damaging, steering behavior down misdirected paths while foreclosing better alternatives.