It's also kind of pointless to talk about how to falsify a class of thing we have affirmative evidence is happening. When we found out about PRISM, the heads of the companies said they didn't know anything about it. So either they were lying or their subordinates successfully kept it a secret from them.
Moreover, the interesting question isn't whether it's happening right now. It's that given we know it can happen, how do we prevent it from happening? For example, by using software with published source code that runs on your own device instead of someone else's.
Can you elaborate on the logic behind this a bit more and perhaps explain how it wouldn't justify McCarthyism as well? (e.g. we know that there was a Soviet spy in one office, so every office that acts displeasingly is filled with Soviet spies)
The problem with McCarthyism is that you're accusing people of being spies without any real evidence. They are, in all likelihood, innocent people being punished for no reason. It becomes a witch hunt and a pretext for punishing anyone you don't like but have no legitimate reason to punish.
By contrast, "spies exist" is a thing that we know is true. There have been documented instances. We don't need to know that there is there is a spy in a particular company at a particular time to know that we should take rational countermeasures against them. Encrypt everything. Eliminate centralization to avoid a single point of compromise for millions of people.
This can't be used for a witch hunt because it's defensive rather than offensive and isn't singling out any particular entity for special scrutiny. If we don't know which particular company or technology is targeted at which time, if we don't know if the adversary is Russia or China or organized crime or corrupt law enforcement, the course of action is still the same. Make mass surveillance as difficult as possible for everybody everywhere.