Almost 100% of the "scientific community of mathematicians" is funded by the government. They can't be trusted either. If they want to publish something that is considered to be a "threat to national security" the agencies have multiple avenues at their disposal to "convince" them to not publish.

> And finally, we wouldn't get laws like this.

Codifying covert practices into law has the big advantage to make the whole oppressive surveillance state much more efficient. Gone are the days of "parallel constructions". Also the chilling effects of total surveillance alone might be enough to prevent the opposition to be effective.