That's fair, but I think if AI companies would be legally required to disclose the sources of their training data and if you make some successor to robot.txt legally binding as well (both is being discussed in the EU for example), at least the "bigger players" in the AI industry would respect the rules. Better than nothing