Nice try ! But the fact that the solution to protecting children comes with the maximum boost of government powers in the online world (across the set of all possible ways to protect children) is not a coincidence.
Isn't it a bit early to draw such conclusions ?
We are just getting started with AI use, especially in tech / engineering teams, and have only scratched the surface with regards to what is possible.
Maybe those people are trying to make the point that if the government can, without public scrutiny, declare any organisation as terrorist and then arrest anyone who wears t-shirts mentioning that organisation - then the government has way too much power.