I think you're suggesting a centralized actor to verify the age of users which is problematic on multiple accounts. Practicality, privacy, enforcement, liability to coercion/corruption, and so on. This might be a wild take, but I think platforms could do a lot more to help with moderation to prevent child predators. Instead of just being okay with them finding outside avenues to converse with their victims through their platform, maybe use those algorithms to detect and prevent it in the first place, whether or not said child predators are whales for your platform or not.

There could be several verifiers. You pick the one you trust the most.

Or you pick three verifiers of your choice who you coordinate between to get verified, without knowing who each other are.