Information topics and means of sharing information that are subject to analysis by a web-of-trust of this type is a very open topic.
Initially, I think it's likely that it could be used to analyze specific domains of information (such as "real identity" or "digital identity" verification).
But looking beyond that, it could also be used to rate the quality of specific posts on Hive, for example. But I should be clear, that is by no means a major focus of this project, it's just an example of how it could be used. Ultimately I think it can be useful for rating most of the information we have, but figuring out how to best get to that point in a practical way isn't completely clear to me yet, it's another research topic in and of itself.
But to answer your question, if we look at that scenario (using it to rate posts on Hive), there's no real way to stop other people from rating your posts using such a system. You wouldn't be able to "opt out", because the analysis takes place on their own computers, although you could register your objections to such rating, I suppose.
Perhaps there would be a way to tag such posts where this rating system was less important. Many genres such as fantasy writing, philosophy, religion, politics, advice, predictions, etc.. These topics were never meant to be argued with scientific facts. The "what if" questions are often the purpose of the writing, to help people explore the possibilities.