Imagine someone uploading photos to a free app and getting a vague notice that their images may train models. What can they actually do with that? The data is already mixed into an aggregate system, so a per-use alert often becomes another ignored checkbox rather than real control. The other side's fairness instinct is understandable, but a blanket right to know every training use creates an impossible accounting burden and false precision. Better to require clear public disclosure of training practices and meaningful opt-outs where feasible, so users understand the rules without pretending each contribution can be tracked.

My opponent treats a knowledge gap as proof of a right to individual notice, but those are different things. The survey shows confusion, not that per-use alerts would fix it. A notice at collection sounds cheap, but a genuine right to know every training use means endless, vague alerts and false precision about data already mixed into an aggregate system. Public disclosure of training practices, plus meaningful opt-outs where feasible, gives people clearer rules and more real leverage than a checkbox they will ignore.

The other side says notice at collection solves this, but that quietly changes the claim. A category notice saying "we may train models on uploads" is exactly the public disclosure I already favor; it is not telling users when their data trains an AI. If the right is truly "when," it either collapses into vague, repeated alerts about data already aggregated, or becomes per-use tracking. The survey shows people feel little control, not that a checkbox fixes it. Clear public rules and real opt-outs where feasible give more leverage than a notice people click past.