I get that question a lot and I can understand it, so there are two things here at work, one is freedom of speech. Depending on your local rules (I am from Germany so our rules i.e. concerning the Nazi past are different than the US rules) freedom of speech should be guaranteed, and opinions shouldn't be labeled as desinformation or manipulation based on their content.

What we are monitoring are deceptive patterns on a text or transcript level. Deceptive patterns can be things like information inconsistency in one post, context shifts in one post that are used to reframe something, or video patterns like fake statistics or fake headlines that are not consistent with the main content.

All of these patterns are actual science backed psychological manipulation patterns and they are consistently used in the most viral posts we detect. My perspective after one year working on this is that the average media literacy is even lower than we think and that we build an evolutionary system with the social media platforms that is optimized to increase the performance of digital manipulation actors.