The article explores concerns about AI agents' impact on children's development of personal accountability. While agents can prevent mistakes and offer convenience, the author argues that children need to make decisions and experience consequences to cultivate judgment. Growing up with agents doing their bidding could hinder the learning process essential for autonomy. The piece suggests that future AI design for children should prioritize opportunities for learning from outcomes, perhaps by allowing children to experience mistakes and then guiding them through reflection, rather than simply optimizing for friction-free outcomes. This approach ensures critical developmental stages are not bypassed.