I would argue, from a layman's perspective, that "I do not know" and "I do not remember" have significant differences, and drawing those parallels would only serve to muddy the waters. In this case it could be possible to show that the testimony was given - even if without other deceit or malice - without competence and "any reasonable person" would be aware of their lack of expertise. Just like it could be possible to show, through evidence such as correspondence, that X had knowledge of Y. I wonder if the system is resilient against such speculated abuse.

Edit: these logical inversions made my ESL head spin a bit. The point was that "I do (not) know" is more readily falsifiable than "I do (not) remember".

In a trial for perjury, the person would be asked, under oath, how they came to believe the information they claim they believed when they gave the testimony that is under dispute. If they actually just made it up or guessed, they would either have to say so, or commit perjury by lying about it. And if the lie was, for example, that so and so gave them a briefing and they understood them to be saying X, then so and so would be called as a witness and their testimony about whether such a briefing took place, and what was said if it did, would be given in evidence.