When you are in a physical space, you can look around you and somewhat reliably observe who can see you, hear you, and observe your actions. You've always then had the choice to base your actions and your speech on who will observe them, with the further knowledge that they might go on to share this with others.
The difference isn't that what you do now can be shared. It's the ease of sharing and the fidelity with which it can be shared.
I mostly agree, but there's part that you're missing. Ease is a critical issue, but the capacity to record was substantially different back then too, and thus the validity of what was being shared. There's a big difference in someone claiming you said something vs someone having a recording of it. 20 years ago everyone wasn't walking around with a camera and microphone in their pocket. It is not only the ease in sharing and identifying, but even the base capacity to perform the action in the first place. If you were being recorded on video in the past, you'd likely notice the bulky camcorder mounted on some guy's shoulder. (not even mentioning that resolution was very different and depth of field is not a negligible).
20 years ago you did not have a reasonable expectation to be recorded in the public physical space. This is contingent upon the probability of someone being able to perform an action, and even more so without you being aware of said action being performed. That wasn't that long ago...
That's what I was getting at with the fidelity attribute.
Before you could observe an event and remember it (poorly) and share it (slowly) but technology, for better or worse, has greatly increased the possible fidelity of that memory, which can now shared endlessly without losing quality.
The difference isn't that what you do now can be shared. It's the ease of sharing and the fidelity with which it can be shared.