On the one hand, I like that it's focused on the broader concept of trust instead of trying to be some kind of AI slop detector. Focusing specifically on AI invites obnoxious debates from the braindead AI bros, so this might make it easier to get adopted by avoiding politics altogether.
... But on the other hand, I could see this going horribly wrong and being abused to bully people. If it becomes common practice to inherit a vouch list from OSS projects, then upstream maintainers essentially become Reddit moderators; Omnipotent grease lords with the power to make anyone's life hell with a single commit.