I think they could instead limit the possible exposure by building tiers/gates a post has to move through to truly go to a mass audience. First it would start with a mass audience of 10^1, mostly direct connections. If no keywords were triggered for misinformation, and none of the eyeballs on that post marked it as misinformation or offensive, then send it to 10^2 people outside their immediate social circle. Repeat until it hits 10^4, 10^5 or whatever and then it hits a tier of truly mass audience. Perhaps here it requires a real review by a moderator if it has keyword matches or isn't from a trusted source. Then it is let out to reach everyone because its just a kitty cat video, or about the best way to stain a table, or all the community content we actually want.
This approach is closer to an A/B test for a social post. If it is posted and highly offensive within a small group of close connections, then it never ripples out of those echo chambers. Again it's about uniqueness of the post, so grandma can click reshare or whatever, when she looks at her history she see's it in her timeline, but it's never massively put on others feed until its gone through above.
Ironically, the social platform killing it right now works somewhat close to this in terms of showing you others posts. My feed shows me a lot of big posters, but like 10% of the time I am reached by people i've never viewed, with a post with like 5-10 views. Then they measure engagement (how long did you watch, did you like, etc), before they decide to push it to more people. (My summary of their algorithm, not theirs). I'm advocating for similar but to gate for offensive/misinformation. Seems like it would work like a sieve where good/unoffensive content would rise to the top "in general" and the bad stuff would stay with a small amount of eyeballs.