Hello HN,<p>we are three researchers (Felix, Jonathan, Johannes) and over the past years, we've been doing research on polarization and misinformation. An earlier project from us is Quality News (<a href="https://news.social-protocols.org/" rel="nofollow">https://news.social-protocols.org/</a>). This is our first prototype for a platform for collaborative fact-checking.<p>What we have so far:<p>- automated detection for rhetorical fallacies in posts. This is done using gpt-4o-mini. We might release that as a library. Let us know if you're interested.<p>On the roadmap:<p>- calculate fact check results taking all the votes and comments in a thread into account<p>- comment ranking by measuring how mind-changing a comment is. We'll decide this by analyzing voting patterns using bayesian statistics.<p>- game theoretical reputation system with incentives for honesty - bias correction for votes with matrix factorization (similar to Community Notes)<p>- crowdfunding of fact checks (money distributed to contributors based on the effectiveness of their comments and votes)<p>We're aware that the true/false vote buttons are not ideal for all the content people will post. But we think it's a start. If you have ideas in this direction, please let us know.<p>We appreciate any feedback!