Crowd-sourced fact-checking tools like X's Community Notes represent a novel approach to combating misinformation, one that relies not on top-down adjudication, but on epistemic consensus among users. Interestingly, fewer than 10% of submitted notes ever become visible — not because they are inaccurate, but because they fail to meet a threshold of cross-ideological agreement. This has led some to think that the policy wrongly prioritizes consensus over truth. But this raises an interesting question: what would an ideal model of crowd-sourced fact-checking look like? As more companies, such as Meta, adopt a Community Notes model, this question deserves sustained scrutiny. This project maps out a vision of how Community Notes should operate.
