I am a professor in the field of cryptography and there are export restrictions here as well. Under no circumstances would I allow the US government to block publication of my papers or talks based on the theory that my speech is “a technology export.” It’s not even clear that this would be allowed under the First Amendment. The government could obviously try to advance that theory, but I strongly suspect they would run into a hailstorm of First Amendment litigation and public blowback.
To me it feels like these prosecutions are an attempt to do exactly that — place restrictions on academic speech and collaboration — but to do it through the “backdoor” of funding agencies and fraud accusations. In other words, if the US said “we are prosecuting you for sharing your open research publications with Chinese scientists,” this would be politically and maybe constitutionally infeasible. So instead they’re trying to achieve effectively the same chilling effect by accusing these scientists of criminal fraud in their funding applications. It is also noteworthy that these cases keep falling apart, which makes me think that they’re exactly as weak as I'm making them out to be.
At the end of the day we need to be very careful about this. We are a country that values free speech, has a constitution that guarantees it, and also needs open speech to make science work. “The government passed some laws restricting technology export” does not actually mean the government can and should be able to restrict scientific speech, and it definitely shouldn’t let them do it through weird indirect means like this.
ETA: It is important to draw a distinction between proprietary or classified IP (and physical equipment) where espionage or theft of trade secret accusations might make sense vs. open scientific research where it sure doesn’t. These cases keep trying to blur the distinction. I suspect if they were clear examples of the former then the cases wouldn’t keep collapsing.
ETA2: I have NSF funding and I just want to add that NSF regulation is byzantine. There are too many regulations surrounding NSF applications and reporting for a human (with an actual job) to be 100% certain that they haven't misstated something or omitted a critical detail. I would guess every US scientist is vulnerable to this sort of (selective, politically-motivated) prosecution.